Recent Strategies to Improve the Photocatalytic Efficiency of TiO₂ for Enhanced Water Splitting to Produce Hydrogen

Abstract

Hydrogen production is one of the best solutions to the growing energy concerns, owing to its clean and sustainable assets. The current review gives an overview of various hydrogen production technologies, highlighting solar water splitting as a promising approach for its sustainable production. Moreover, it gives a detailed mechanism of the water-splitting reaction and describes the significance of titania-based catalysts for solar water splitting. It further highlights diversified strategies to improve the catalytic efficiency of TiO₂ for the enhanced hydrogen production. These strategies include the doping of TiO₂, dye sensitization, and the addition of co-catalysts. Doping reduces the bandgap by generating new energy levels in TiO₂ and encourages visible-light absorption. Sensitization with dyes tunes the electronic states, which in turn broadens the light-absorption capacity of titania. Constructing heterojunctions reduces the charge recombination of TiO₂, while co-catalysts increase the number of active sites for an enhanced reaction rate. Thus, every modification strategy has a positive impact on the stability and photocatalytic efficiency of TiO₂ for improved water splitting. Lastly, this review provides a comprehensive description and future outlook for developing efficient catalysts to enhance the hydrogen production rate, thereby fulfilling the energy needs of the industrial sector.

1. Introduction

Due to greenhouse gases, climate change, decreasing fossil fuels, and increasing energy requirements, we will need a transition from a society that depends on non-renewable fuels to one that is more reliant on renewable energy in the upcoming era [1]. It is alluring to search for an alternate energy source that can replace conventional fossil fuels while not polluting the environment. Although photovoltaic cells reveal an ideal solution to the energy problem by offering a carbon-free solution, they lack a cost-effective and efficient energy-storing process that could supply energy at nighttime [2]. On the other hand, converting solar fuel into its chemical form can be a better approach for solving energy problems. In this regard, hydrogen (H₂) is an excellent source of clean, highly combustible, and renewable energy. H₂ energy can be produced in a variety of ways (Figure 1) [3], can be stored in different forms (gaseous, liquid, and metal hydrides), and can be transformed into a variety of fuels. The ecological effects of the energy zone are reduced because no carbon is emitted when H₂ is produced from water (H₂O) or any other renewable resource [4,5].

2. Hydrogen Production Methods

H₂ can be produced through a variety of methods, such as biological, thermochemical, steam reforming, and coal gasification [3]. Since methane gas has the largest H₂/C ratio when compared to other hydrocarbons, it has a lower risk of byproduct formation. Thus, methane is mostly used in steam reforming for H₂ production. Different steps involved in the steam reforming process to produce H₂ are as follows: (i) preferential oxidation reaction (POR); (ii) steam reforming; (iii) high-temperature water gas shift reaction (HT-WGSR); and (iv) low-temperature GSR (LT-WGSR) [6]. Another method for producing H₂ is coal gasification, which involves partially oxidizing coal at high temperature and pressure (around 5 × 10⁶ Pa). This oxidation produces a mixture of CO, CO₂, H₂, and other compounds. A mixture of CO and H₂ is left only when the temperature is raised to 1273 K [7]. Biomass, i.e., animal and plant waste, is utilized in thermochemical and biological processes to produce H₂. Examples of thermochemical H₂ generation processes include gasification and pyrolysis, whereas biological H₂ production processes include fermentation, biophotolysis, and the gas shift process [8].

Renewable energy sources include hydropower, geothermal energy, wind energy, and solar energy [9]. However, the development of an energy fountain is hampered by the direction of the wind, a shortage of storage capacity, and the construction of larger dams. H₂ is generally considered a clean energy source, but this is not always the case. Carbon dioxide, a greenhouse gas that contributes to environmental damage, is produced during its production from coal, natural gas, or biomass. Therefore, there is a need for an alternative H₂ source. Production of H₂ from water splitting is a carbon-free solution, so this process is an ecofriendly way to address the current environmental issues. Water splitting can be categorized into three types: photobiological, thermochemical, and photocatalytic. Water and organic biophotolysis are two categories for the photobiological splitting of H₂O, commonly known as biophotolysis [4]. H₂O biophotolysis is the mechanism by which sunlight produces H₂ through the water-splitting action of green algae or cyanobacteria (oxygenic bacteria). Enzymes such as hydrogenase and nitrogenase are responsible for this water splitting. Despite being a clean method of producing H₂, there are a few known drawbacks, such as a low output and the enzyme’s potential for toxicity. One more reason that makes this process problematic is the requirement of a bioreactor, which is challenging to design [3]. Biophotolysis is a process by which organic waste is decomposed in the presence of aerobic bacteria to produce H₂. Although this approach yields higher results, it is problematic due to CO₂ emissions [10]. By using solar reactors, heat from sunlight is concentrated to achieve a temperature of 2273 K, which is then used to split H₂O with the help of a catalyst (such as ZnO). Although it is a facile approach, it has several disadvantages. First of all, managing heat is really difficult. Secondly, it is very costly since large-scale solar concentrators are widely sought to achieve high temperatures [11].

By utilizing solar light, the photocatalytic splitting of H₂O (Figure 3) is the most efficient method of producing H₂ [12] without any greenhouse gas emission. Since solar energy, which has a power density of 400–1000 W/m², is another renewable energy source [13], it can be converted in a variety of ways to produce heat or electricity. Moreover, the quantity of radiation that falls on earth in just thirty minutes can meet the world’s annual energy needs [14]. Solar photocatalytic water splitting has several advantages, such as the highest energy conversion efficiency, independent O₂ and H₂ emissions, and the least expensive and most environmentally beneficial approach. Reactors can be constructed on a small scale, increasing their market efficiency and lowering the need for solar power systems.

Table 1. Hydrogen production methods [15].

Method	Technique Used	Feed Stock	Title
Photonic	PV electrolysis	Water	Electricity is produced using PV panels
	Photocatalysis	Water	Using electron–hole pairs produced by the photocatalyst, water is divided into O2 and H2
	Photo-electrochemical method	Water	A hybrid cell generates voltage and current to absorb light
Biochemical	Dark fermentation	Biomass	A hybrid cell generates voltage and current to absorb light
Electrical + Thermal	High temperature electrolysis	Water	Water splitting occurs at a high temperature using thermal and electrical energy
	Hybrid thermochemical cycle	Water	Cycles of chemical reactions are driven by electrical and thermal energy
	Coal gasification	Coal	Coal is transformed into syn-gas
	Reforming of fossil fuel	Fossil fuel	Fossil fuels produce H2 and CO2
Electrical	Electrolysis	Water	Water is split by direct current into O2 and H2
	Plasma	Fossil fuels	Carbon soot and H2 are produced when natural gas is pushed through a plasma arc
Thermal	Thermolysis	H2S	H2S thermally decomposes at high temperatures
Thermochemical processes	Water splitting	Water	Water is oxidized by converting sunlight into electron–hole pairs
	Biomass conversion	Biomass	Conversion of fermentable hydrogen via thermo-catalysis using biomass-based materials
	Gasification	Biomass	Biomass is transformed into syn-gas
	Reforming	Biomass	liquid bio fuels to transform into H2

below provides a list of a few primary techniques used for H₂ production in the present era [15].

Among several possible techniques, water electrolysis driven by solar radiation is a significant choice. It can be achieved by photoelectrochemical cells or by self-supported systems.

3. Mechanism of Photocatalytic Water Splitting

When light falls on a photocatalyst, a chemical reaction occurs that produces hydrogen from water [16]. Three steps make up the total photocatalytic water-splitting reaction. (i) The capture of energy above the bandgap of the semiconducting material leads to the formation of electron–hole pairs within the valence band (VB) and conduction band (CB), respectively. (ii) The holes (h⁺) oxidize the water, releasing H⁺ and O₂. In order to create H₂ by the reduction of h⁺ ions, photogenerated electrons (e⁻) are then transferred towards the conduction band (CB). (iii) The reduction of H⁺ ions occur to form H₂. Understanding heterogeneous photocatalytic water splitting requires an understanding of the sacrificial donor and the catalyst [17].

If a catalyst absorbs a photon of suitable energy which is equal to its bandgap, then it can split water but the efficiency of the process is limited because of the lower hydrogen yield which is ascribed to low energy absorption or charge recombination on the catalytic surface. To reduce the rate of charge recombination, a sacrificial donor is employed to increase the yield of H₂ [18]. The segregation of charges requires the co-catalyst/buffer layers to be combined with the electrodes and the photocatalyst. Surface oxidation, exposed surface area, and ion absorption are all factors affecting the bandgap potential which is a key criterion for enhanced water splitting. Figure 2 shows the mechanism of photocatalytic water splitting.

The general water-splitting reactions are

H₂O + 2h → 2H + 1/2O₂, E° _oxidation = −1.23 V

(1)

2H⁺ + 2e⁻ → H₂, E° _oxidation = 0.00 V

(2)

A minimal thermodynamic potential of 1.23 eV is required for photochemical water splitting [19]. The following are the major obstacles in photocatalytic total water splitting: (i) A large number of photocatalysts lack a suitable band alignment which is needed for overall water splitting. (ii) The most difficult step in the photocatalytic splitting of water is thought to be the O₂ evolution half reaction, due to the four-electron transfer process coupled with the two-proton removal from H₂O to form an O-O bond, which results in mismatched stoichiometric rates [20]. (iii) The generated H₂ and O₂ quickly react back to form water again on the catalytic surface, especially on non-noble metallic surfaces, which results in a low energy conversion efficiency. Currently, the maximum solar-to-hydrogen conversion capacity of CdS-based catalysts is 1.63%, which is far from real application requirements [21].

Figure 2. The stepwise mechanism for photocatalytic water splitting. Reprinted with permission from Ref. [22]. Copyright 2021, Elsevier.

Figure 2. The stepwise mechanism for photocatalytic water splitting. Reprinted with permission from Ref. [22]. Copyright 2021, Elsevier.

3.1. Types of Reactions

In photocatalytic water-splitting processes, H₂ can be produced in two different ways:

• Photochemical reactions;
• Photoelectrochemical reactions.

3.1.1. Photochemical Reactions

The reactions which are conducted in photocatalytic suspensions in water are termed photochemical reactions, and the process is conducted by the immediate use of energy to cause a chemical change [23]. After absorbing enough energy, the splitting of water with a suitable photocatalyst results in the evolution of both H₂ and O₂. It becomes difficult to regulate a photocatalyst’s specific characteristics, like its capacity for light absorption when suspended, changes in the reaction’s pH, and the substrate’s concentration. The H₂ gas evolved at the semiconductor–electrolyte junction serves as an interface to create the voltage required for water splitting. However, it is quite challenging to analyze the experimental results since these factors influence the reaction kinetics, making it rather arduous. We have shown that an artificial photosynthesis system based on photochemical diodes is capable of effective, unaided overall water splitting without the need for a sacrificial reagent [22]. A photochemical diode array demonstrated a solar-to-hydrogen efficiency of 3.3% in neutral conditions (pH 7.0), in which the total water-splitting process was carefully regulated by regulating charge carrier movement at the nanoscale [24].

3.1.2. Photoelectrochemical (PEC) Reactions

PEC water splitting is a potential remedy for the present world energy consumption and environmental issues that converts solar radiation to H₂ fuel. The water oxidation reaction, which uses four electrons, limits the total efficiency of water splitting. The ideal photocatalyst for water splitting in PEC ought to be very selective, must have a narrow bandgap, and be highly active, durable, and cheap [25]. A conventional PEC setup involves an electrolyte, a photoanode, or a photocathode, sometimes both. Typically, a there is light response in an n-type semiconductor which serves as the foundation for the photoanode; it has the ability to produce electron–hole pairs when stimulated by light. In an external circuit, electrons produced in the photoanode flow to the photocathode, where they are accepted by hydrogen ions contained in the electrolyte solution to make H₂. This reaction can be written as follows [26].

Photocathode: 2H⁺ + 2e⁻ → H₂

(3)

Photoanode: 2H₂O + 4H⁺ → O₂ + 4H⁺

(4)

The overall reaction equation for splitting water is

2H₂O → O₂ + 2H₂

(5)

PEC is regarded as a promising approach for H₂ production to meet the energy requirements of the industrial circular system [27]. An electric current flowing through an external circuit drives the reaction in PEC. The photocatalyst is deposited to create a photoanode to carry out the reaction of water splitting in solution. The photon’s energy can be converted into electrochemical energy from solar radiation falling on the PEC surface, which causes H₂O to split into O₂ and H₂. When electrons fall on the photocathode, H₂ is created there. The scattered light is subsequently transformed into a storable form [28]. To increase the photo electrocatalytic activity of titania-based PEC cells, two methods have been presented. The first method, known as photoinduced reduction processes at the TiO₂ surface, entails altering the pristine TiO₂’s structure by chemical doping, surface/subsurface defects, changes in surface morphology, and particle size, all of which result in a lowering of the bandgap. The second method, referred to as “photo-induced oxidation processes at the TiO₂ surface”, relies on using oxidizing substances like various alcohols that are used as an electron donors in PEC cells [29].

Working Principle of Photoelectrochemical (PEC) Reaction

The fundamental process in a photoelectrochemical (PEC) cell involves water splitting by applying an external bias to the photoelectrodes in order to convert sunlight into hydrogen fuel by using an electrolyte. PEC water splitting has four fundamental steps: Sunlight strikes the photoanode first, producing electron–hole pairs. Secondly, water oxidation is caused by photogenerated pores at the activated photoanode surface. Thirdly, the electrons produced by the incident photons move from the photoanode to the photocathode via an external cable, where they are reduced by the electrons to generate hydrogen gas at the photocathode surface [30]. The PEC cell’s photoanode is formed of an n-type semiconductor, while the photocathode is composed of a dark metal electrode. These two electrodes are separated by an ion-exchange membrane as shown in Figure 3. Whenever the semiconductors are exposed to sunlight, the n-type material is used in the photoanode to generate holes for the oxygen evolution process (OER) (Equation (6)), while electrons move to the photocathode for hydrogen evolution reaction (HER) (Equation (7)). To create a tandem cell, certain p-type materials have been created. The former is a four-electron reaction, whereas the latter is a two-electron reaction, making the former far more challenging. Equation (8) depicts the entire water-splitting reaction, which is a highly energetic reaction with a Gibbs-free energy of ∆G° = 237.2 kJ mol⁻¹ per mole of produced H₂ (using the Nernst equation as a reference, ∆E = 1.23 V) [31].

Reaction at anode

2H₂O + 4h⁺ → O₂ + 4H⁺ E°_O2/H2O = 1.23 V vs. RHE

(6)

Reaction at cathode

2H⁺ + 2e⁻ → H₂ E°_H+/H2 = 0 V vs. RHE

(7)

Net reaction

2H₂O → 2H₂ + O₂ ∆E = 1.23 V

(8)

However, as shown in Figure 4, the upper section of the valence band should be at a greater positive potential compared to the water oxidation potential into oxygen (1.23 V at pH = 0), in order to improve PEC’s water-splitting efficiency [30].

4. Desired Characteristic of the Material Used for PEC

Selecting a good photocathode and/or photoanode material is the most important step in PEC water splitting. Semiconductor materials with the following properties are necessary for an effective PEC water splitting which could serve as a photoanode or a photocathode.

4.1. Adequate Band Locations and Bandgap Energy

In total, 52% of the radiation of the solar spectrum lies in the infrared region (700–2500 nm), 43% is visible (400–700 nm), and 5% is UV (300–400 nm). Hence, to enhance the efficiency of the catalyst, a significant amount of visible light must be absorbed, which in turn depends on the semiconductor’s bandgap. Given that the O₂/H₂O potential is 1.23 V vs. NHE (pH = 0) and the proton reduction potential is 0 V vs. NHE, incident photons with a minimum energy of 1.23 eV, or roughly 1100 nm in wavelength of light, are needed to achieve the predicted minimum bandgap for water splitting. However, a minimum bandgap of ~1.8 eV is required, equivalent to light absorption at around 700 nm, when taking into account the thermodynamic energy losses (0.3–0.4 eV) occurring during charge carrier transit and the overpotential requirement for appropriate surface reaction kinetics (0.4–0.6 eV). The solar spectrum indicates that the maximum bandgap energy is 3.2 eV due to a sharp decrease in solar intensity below 390 nm.

A single semiconductor photoelectrode should have a bandgap energy of 1.9 eV to 3.2 eV, in order to obtain a significant photovoltage. A bandgap of 2.0 eV is theoretically required for optimal sunlight usage, in addition to the thermodynamic band position requirements [32]. In actuality, though, it has rarely been realized because the anticipated photovoltaic output is rarely reached. Even with the maximum current density, unbiased water splitting cannot be solved without a high photovoltage value (>1.61 V). Therefore, external bias or tandem devices are required to supply the additional voltage required to split water. The bandgaps and band locations of common n-type and p-type semiconductors used for PEC water splitting at pH = 0 are shown in Figure 5.

Since VH (as illustrated in Figure 6) relies on the pH of the solution (by −0.059 V per pH unit) with regard to the redox potential of the electrolyte, it is noteworthy that the band locations of oxide semiconductors fluctuate depending on the electrolyte’s pH. Moreover, the water reduction and oxidation potentials alter in a way that is comparable to the band positions (i.e., −0.059 V per unit pH), as per the Nernst equation. The broad consensus is that H⁺ and OH⁻ are potential-determining ions (PDIs) deposited on the semiconductor surface inside the Helmholtz layer, as indicated by the Nernstian dependence. Other compounds, including metal sulfides, may have varying PDIs, which makes the pH dependence more complex [33]. While the band locations are constant with regard to the water redox potentials, most semiconductors, and oxides in particular, exhibit the same pH dependency as shown by the Nernst equation. As a result, the relative difference between an oxide photocatalyst’s band locations and the water redox potential does not depend upon pH.

Figure 5. Band locations of different semiconductors in relation to the water-splitting redox potentials at pH = 0. Reprinted with permission from Ref. [34]. Copyright 2017, Royal society of chemistry.

Figure 5. Band locations of different semiconductors in relation to the water-splitting redox potentials at pH = 0. Reprinted with permission from Ref. [34]. Copyright 2017, Royal society of chemistry.

4.2. Effective Separation and Movement of Charge Carriers within the Semiconductor

Low STH efficiencies are largely caused by fast charge recombination, so an effective charge carrier separation and transport strategy is needed. This strategy depends on the material’s inherent characteristics, such as hole and electron mobility, as well as its extrinsic characteristics, such as crystallinity and nanostructure.

4.3. Stable and High Catalytic Activity

Electron–hole recombination would normally result from surface charge accumulation, which can be prevented by appropriately fast surface reaction rates. For many potential water-splitting semiconductors, especially metal sulfides, photocorrosion is a significant issue since it happens when the holes and electrons produced by photons break down the photocatalyst rather than oxidizing or reducing the water. The relative locations of the semiconductor band edges and their corresponding decomposition potentials determine these photocorrosion reactions. If the anodic decomposition potential (Epd) of the semiconductor is higher than its valence band potential, anodic photocorrosion may happen. On the other hand, if the cathodic decomposition potential (End) is lower than the semiconductor’s conduction band, cathodic photocorrosion may happen. The actual values of the decomposition potentials depend on the pH of the electrolyte; therefore, metal oxides like ZnO and BiVO₄ or metal sulfides like CdS and MoS₂ can easily undergo anodic photocorrosion. Nonetheless, because of their extremely slow decomposition reaction kinetics, typical photoanode materials like Fe₂O₃ and TiO₂ are thermodynamically stable despite having an anodic deposition potential above the valence band potential.

For practical applications, the photoelectrode materials should also be inexpensive and made of elements that are abundant on Earth. This is essential to supporting the case for a cost-effective scaling up of solar-to-fuel devices [34].

• Photocathodic materials

In PEC water splitting, the photocathode typically consists of a p-type semiconductor deposited on conductive substrates. Water is oxidized on the counter photoelectrode and reduced on the semiconductor surface when exposed to light, as depicted in Figure 7. In order to produce hydrogen, the photocatalysts’ conduction band edge needs to be more negative than the potential for hydrogen evolution from the perspective of electrochemical potential. There are a lower number of reports based on p-type semiconductors as photocathodes for PEC water splitting as compared to n-type semiconductor photoanodes [35]. Cu₂O is a common p-type semiconductor employed as a photoelectrode for PEC water splitting, while it quickly degrades owing to self-reduction by photogenerated electrons [35]. In order to decrease the self-reduction of Cu₂O and improve the PEC characteristics, WO₃/Cu₂O p-n junctions have been created. To enhance their PEC performance, p-n heterojunctions have also been produced. As illustrated in Figure 8, Colleen at al. [36] synthesized p-n Cu₂O homojunction solar cells by electrochemically installing an n-Cu₂O layer on a p-Cu₂O layer. The produced p-Cu₂O and n-Cu₂O layers’ inherent doping levels were very low, and they made ohmic junctions with Cu metal. Compared to other p-n homojunctions, the greatest cell performance (a η of 1.06%, a V_OC of 6.21 V an I_SC of 4.07 mA cm⁻², and a fill factor of 42%) was achieved. Paracchino et al. reported that the Cu₂O photocathode, depicted in Figure 9, demonstrated photocurrents of up to −7.6 mA cm⁻² at a potential of 0 V versus the reversible hydrogen electrode at moderate pH [37].

In another study, the Cu₂O photocathode was activated for hydrogen evolution with electrodeposited Pt NPs and protected against photocathodic decomposition in water by nanolayers of Al-doped zinc oxide and titanium dioxide. A Pt-modified Si photovoltaic cell (Pt/SiPVC) has also been reported as an efficient photocathode for PEC water splitting that uses p-n radial junctions with a p-type Si substrate, and B-doped Si with the decoration of Pt has frequently been used as a p-type photocathode for PEC water splitting [38]. The cheap and plentiful Si supply makes Si-based photoelectrodes for PEC water splitting a promising option; however, their efficiency still needs to be increased.

Figure 7. Semiconductors installed on a substrate as a photoanode for PEC water splitting. Reprinted with permission from Ref. [39]. Copyright 2014, Royal society of chemistry.

Figure 7. Semiconductors installed on a substrate as a photoanode for PEC water splitting. Reprinted with permission from Ref. [39]. Copyright 2014, Royal society of chemistry.

Figure 8. SEM image of a p–n Cu₂O homojunction. Reprinted with permission from Ref. [36]. Copyright 2012, Royal society of chemistry.

Figure 8. SEM image of a p–n Cu₂O homojunction. Reprinted with permission from Ref. [36]. Copyright 2012, Royal society of chemistry.

Figure 9. Schematic illustration of the electrode structure of the surface-protected Cu₂O electrode. Reprinted with permission from Ref. [37]. Copyright 2011, Nature.

Figure 9. Schematic illustration of the electrode structure of the surface-protected Cu₂O electrode. Reprinted with permission from Ref. [37]. Copyright 2011, Nature.

• Photoanodic materials

Typically, n-type semiconductors installed on conductive substrates make up the photoanode in the PEC water-splitting arrangement [39]. Under light illumination, as seen in Figure 9, electrons are transported to a counter electrode via an external circuit, while photoexcited holes build up on the surface of the photoanode semiconductors and are consumed in oxidation reactions. In order for the photoanode to produce oxygen, the photocatalysts’ valence band edge needs to be more positive than the potential for the evolution of oxygen from the perspective of electrochemical potential. One of the main benefits of PEC is that an external voltage bias can be used to make up for any inefficiencies and quicken the excited charge carrier separation process. However, zero bias is preferred once the PEC systems are well-aligned with a suitable semiconductor. This section begins with the conventional TiO₂ photoanode and discusses the latest developments in TiO₂ and TiO₂-based hybrid photoanode systems (Figure 10).

Because of its lower cost and superior chemical and optical stability, titanium dioxide (TiO₂) is the most frequently used n-type semiconductor for PEC water splitting [41]. Because of their inherent wide surface area and unidirectional charge flow, 1D titanium dioxide nanotube arrays (TNAs) have drawn a lot of interest among the diverse nanostructured TiO₂ [42]. TNA has a significant deal of promise for large-scale applications and is easily manufactured by anodizing metal titanium foil or plate [43]. Via a mild hydrothermal reaction, 1D TiO₂ nanowire arrays have also been coated on glass coated with a transparent conducting oxide (TCO) [44]. When using 1D nanowire arrays, N719 dye-sensitized solar cells can achieve photoconversion efficiencies of up to 5.02%, which is significantly greater than when using ordinary TiO₂ powders. Since titanium dioxide has a broad bandgap of approximately 3.2 eV and is not sensitive to visible light, it exhibits a modest photoelectrochemical activity [45]. Researchers have devised a number of techniques to improve the absorption of visible light, such as doping TiO₂ with metal or non-metal, creating heterojunctions, hydrogenation, or the formation of structural vacancies [46]. In the early 21st century, doping with metal or non-metal (C, N, S, B, etc.) was introduced as a practical way to reduce the bandgap of TiO₂ and increase PEC efficiency [47]. The CB and VB of the acceptor can grow with metal and non-metal doping, respectively. This narrows the photocatalysts’ bandgap and enhances visible-light absorption. Park et al. synthesized high-aspect-ratio carbon-doped TiO₂ nanotube arrays. Under white light illumination, the total photocurrent was over 20 times higher than that of a P-25 nanoparticulate film [48].

Another viable method for harvesting visible light is to modify a TiO₂ nano-semiconductor by adding a second nano-semiconductor with a lower bandgap, creating a heterojunction. In addition to being a photosensitizer, the second nano-semiconductor creates an internal electric field across the interface. Recombination is decreased as a result of the internal potential bias, which significantly facilitates the separation and passage of excited electrons and holes over the dual photocatalyst interface. For example, p-type Cu₂O and CdTe semiconductors have always been used to modify TiO₂ nanotube arrays (TNAs) [49]. TiO₂ is a p-type SC, while Cu₂O and CdTe are p-type SCs. As a result, between them, p–n junctions may form, which helps to separate the excited holes and electrons.

Due to bandgap overlapping between the two distinct photocatalysts, certain heterojunctions can improve photoelectrochemical characteristics and promote charge carrier separation and transfer. The examples shown in Figure 11 and Figure 12 are typical examples. Since catalyst A (Bi₂WO₆) has a higher negative conduction band (CB) than catalyst B (TiO₂), excited electrons from catalyst B (TiO₂) can move swiftly to catalyst A (Bi₂WO₆). In comparison to catalyst B (TiO₂), catalyst A’s (Bi₂ WO₆) valence band (VB) is more positive, and catalyst A’s (Bi₂WO₆) excited holes can move rapidly to catalyst B (TiO₂). For effective water splitting, the excited holes and electrons can thus be swiftly separated and transported.

5. Electrolysis

A scalable, adaptable, and distributed method of producing hydrogen is water electrolysis. Electrolysis currently produces about 4% of the hydrogen in the world. Electrolysis-based hydrogen production can also serve as a buffer which helps to balance the supply and demand of electricity [51]. There are numerous ways to produce hydrogen. Because of its high purity, ease of use, and lack of contamination, water electrolysis is the most widely utilized method for the conversion of renewable energy into hydrogen fuel, yet it holds a minuscule presence in the market [52]. The most basic industrial method currently available for creating almost pure hydrogen is water electrolysis, and in the years to come, its significance is expected to increase. Water electrolysis is based on the movement of electrons, which is regulated by an external circuit. Solid oxide electrolyzers, alkaline membranes, and polymer electrolyzers are the three primary electrochemical technologies for producing hydrogen [53]. Electrolysis is the most effective technique for splitting water, which is a tried-and-true procedure. But since the process is endothermic, the required energy input is provided by electricity. As shown in Figure 13 below, a basic electrolysis unit (an electrolyzer) comprises a cathode as well as an anode submerged in an electrolyte. Typically, when an electric current is given, water splits and oxygen is released from the anode side while hydrogen is produced at the cathode through the following reaction [54].

The following are the overall anode and cathode reactions:

Reaction at anode

4OH⁻ → O₂ + 2H₂O

(9)

Reaction at cathode

2H₂O + 2e⁻ → H₂ + 2OH⁻

(10)

Net reaction

H₂O → H₂ + 1/2O₂ ∆H = −288 kJ mol⁻¹

(11)

This method has the advantage of producing pure hydrogen, as it is a sulphur- and carbon-free technique [55]. However, compared to systems based on fossil fuels, electrolysis has several disadvantages, such as higher cost and energy requirements [37]. Nevertheless, electrolysis is still regarded as a potentially economical method of producing hydrogen locally due to its compactness and small-scale uses. Additionally, it can be used in existing gas stations [56]. It is crucial to maintain a balance between byproduct oxygen and oxygen demand. If the latter for hydrogen generation by water electrolysis does not exceed the potential supply of the former, a significant amount of the byproduct oxygen should be lost. The industrial gas oxygen itself is a crucial component of numerous industries, including glass melting, electric furnaces, and blast furnaces. Thus, by selling the byproduct oxygen to these industries, the minimal expense of creating hydrogen by protein exchange membrane (PEM) electrolysis will be required. On the other side, if a significant oxygen consumer uses PEM electrolysis to produce oxygen, it can sell the hydrogen which makes it a significant choice for water splitting [57].

5.1. Mechanism of Water Splitting in Alkaline Solutions

In alkaline solutions, HER activity is 2–3 times lower compared to acidic solutions and the process is highly dependent on the structure of the catalytic surface. Theoretical studies have shown that HER performance is related to H₂ adsorption in acidic solutions and it is comprised of Volmer/Tafel or Volmer/Herovsky steps as shown below:

Volmer step: H⁺ + e⁻ → H_ad

(12)

Heyrovsky step: H⁺ + e⁻ + H_ad → H₂

(13)

Tafel step: 2H_ad → H₂

(14)

In alkaline solutions, it is mandatory to trade off hydroxyl and hydrogen adsorption (H_ad and OH_ad) along with water dissociation. The Volmer and Heyrovsky steps are demonstrated with water rather than protons acting as reactants.

Volmer step: H₂O + e⁻ → OH⁻ + H_ad

(15)

Heyrovsky step: H₂O + e⁻ → OH⁻ + H₂

(16)

Thus, there is a need to break the strong bonding of H-O-H in alkaline solutions compared to weak bonding in the H₃O⁺ ion. The rational design of catalysts with a higher ability to split water can enhance alkaline HER activity. There are four major factors which may affect the alkaline HER activity that primarily depends upon which reaction pathway (Heyrovsky or Volmer) is followed. These factors include the adsorption of water on catalytic surfaces, water-splitting magnitude, hydrogen binding energy, and the adsorption strength of OH groups. In HER, water adsorption is the first step in alkaline medium which is very low compared to H₃O⁺ adsorption in acidic medium. Hence, developing stronger bonding between the metal and water could be a key solution to improve HER performance [58].

5.2. Ion-Exchange Membrane

To develop efficient, robust, and safe PEC water-splitting systems, lowering cell voltage and improving product separation are essential elements. A robust method that can prevent diffusive and convective crossover of O₂ and H₂, for the practical deployment of large-scale PEC water-splitting systems, is critically important to ensure the system stays below the flammability limit of 4% H₂ in O₂ mixture in a dynamic operation [59]. The low operating current densities of PEC water-splitting systems relative to water electrolysis systems have stricter requirements for product separations [60]. Ion-exchange membranes with a suitable permeability and thickness can provide product separation and minimize crossovers [61]. This separation barrier, however, must be ionically conductive, as to maintain charge balance across the cell. It is also crucial to improve the efficiency and economic feasibility of PEC systems by lowering the overall cell voltage. The use of membranes in the system for product separation adds to the overall cell voltage, therefore, increasing the conductivity of these membranes is key to their development [59]. As described further below, the use of ion-exchange membranes can also lead to improved catalytic performance, further reducing the cell voltage and creating ideal environments for Earth-abundant catalysts for both the oxygen evolution reaction (OER) and the hydrogen evolution reaction (HER) [62].

Ion-exchange membranes, also known as anion-exchange membranes (AEMs) or cation-exchange membranes (CEMs), are made of polymers with ionizable functional groups that enable the selective transport of anions and cations [63]. In PEC systems, monopolar membranes (CEMs or AEMs) are employed to transfer ionic currents across electrochemical cell chambers and block H₂ and O₂ crossing [59]. A water dissociation (WD) catalyst is often positioned at the intersection of the cation-exchange layer (CEL) and anion-exchange layer (AEL) in bipolar membranes (BPMs). At the intersection of the CEL and AEL, there is an increased WD under reverse bias as a result of a catalyst and a strong electric field. Then, distinct acidic and basic streams are created on either side of the BPM by H⁺ ions moving preferentially through the CEL and OH⁻ ions traveling selectively through the AEL [64]. BPMs enable the maintenance of significant pH gradients (0–14) between the cathode and anode chambers in PEC systems. Therefore, for Earth-abundant catalysts for the HER, which is kinetically more possible under acidic conditions, and the OER, which is kinetically more feasible under alkaline conditions, the pH of the catholyte and anolyte can be independently tuned [64,65]. To quantify the ion-transport parameters in these CEMs, AEMs, and BPMs, standard benchmarking techniques must be created in order to enhance ion-exchange membrane technology.

6. Hydrogen Generation Electrolyzers

Electrolysis is a chemical reaction in which a substance is broken down by an electric current flowing through it. Electrolyzers are tools that are essential to this process. Depending on the type of electrolyzer and compounds being processed, the primary purpose of an electrolyzer is to aid in the electrolysis process and generate a variety of chemicals.

A general schematic for producing hydrogen energy using electrolyzers is shown in Figure 14. In short, hydrogen-generating electrolyzers are machines in which the electrolysis process is used to split H₂O into H₂ and O₂ with the help of electricity. Water electrolysis is the name given to this process, which uses two electrodes (an anode and a cathode) and an electrolyte (typically an alkaline solution or a polymer electrolyte membrane). Due to the external electrical potential applied to the electrodes that are dipped in water, hydrogen is generated at the cathode and oxygen is generated at the anodes. After being separated, the hydrogen gas can be gathered, kept, and used as a clean energy source [66]. The establishment of electrolyzers for hydrogen production is important for the storage of renewable energy in hydrogen fuel. Several programs such as power-to-gas and power-to-X are dedicated to store and use hydrogen as a feasible and environmentally responsible energy source [67]. Water electrolysis can also be made more affordable by utilizing renewable electricity from sources like solar cells [68].

Types of Electrolyzers

Although the basic idea behind water electrolysis is simple, it allows for the freedom to create several iterations by utilizing a variety of physiochemical and electrochemical parameters. There are various kinds of electrolyzers, and each has unique benefits and uses [69]. Four main technologies are widely used to classify electrolyzers, and these can be distinguished by the particular electrolyte and operating temperature of each technology. The selection of components and materials is heavily influenced by these distinctions. Alkaline, solid oxide electrolyzer cells (SOECs), proton-exchange membrane (PEMs), and anion-exchange membrane (AEMs) are the four main types of electrolyzers [70].

Table 2. Comparison analysis of electrolysis technologies, including PEM, AEM, alkaline, and SOEC, based on distinct characteristics and attributes [71].

Characteristic	PEM	AEM	Alkaline	SOEC
Electrolyte	Proton-exchange membrane	Anion-exchange membrane	Liquid KOH solution	Solid oxide ceramics (e.g., YSZ)
Electrolyte Material	Acidic/solid (polymer)	Alkaline/solid (polymer)	Alkaline/liquid	O2 or H+ conducting/solid (ceramic)
Working Principle
Operating Temperature	50–80 °C	40–90 °C	60–80 °C	700–1000 °C
Electrode Material	Platinum group metals (PGMs)	Non-precious metals	Nickel-based	Perovskites, ceramics
Efficiency	60–70%	50–60%	65–70%	80–90%
Durability	High	Moderate	High	Low to moderate
Water Quality	High-purity water required	High-purity water required	Low-purity water required	High-purity water required
Gas Purity	High	Moderate to high	Moderate	High
System Complexity	High	Moderate	Low	High
Start-up Time	Short	Short	Moderate	Long
Cost	High	Moderate	Low	High
Scalability	High	High	Moderate to high	Low to moderate
Commercial Maturity	High	Emerging	Mature	Research and development
Applications	Mobility, small-scale hydrogen	Industrial, potential for mobility	Large-scale industrial	High-temperature industrial process

presents an overall comparison of all electrolytic cell types that are commercially available.

The most common type of electrolysis is called alkaline water electrolysis (AWE), which uses high concentrations of KOH as the electrolyte and operates at low temperatures (60–80 °C) [72]. Furthermore, these technologies allow the use of inexpensive and non-noble materials [73]. Newer technologies such as PEM and SOECs can manufacture green hydrogen from renewable energy sources with zero carbon emission and great efficiency [70,74].

One promising path towards large-scale, renewable energy-based hydrogen production is by AEM electrolysis. However, difficulties with membrane electrode installation and reaction kinetics are primarily responsible for the efficiency of AEM electrolysis. Through focused efforts, researchers are actively working to improve AEM electrolysis performance in order to address these constraints [75].

Water electrolysis which includes the AWE, AEM, PEM, and SOEC methods offers a viable pathway for producing hydrogen with a variety of uses in the fields of catalysis and renewable energy [70,75]. However, “gray hydrogen” sourced from fossil fuels dominates the current global hydrogen energy environment. This path diverges from the initial objective of attaining net-zero emissions [76]. Because of this, current research and development initiatives are focused on producing green hydrogen in an eco-friendly manner while also trying to improve the effectiveness and performance of these electrolysis techniques for hydrogen generation at a large scale.

A water electrolysis cell’s working concept is centered on two electrodes separated by an electrolyte. The migration of generated chemical charges, such as cations (+) and anions (−), between the electrodes is facilitated by this medium. Potassium hydroxide (KOH) in a highly concentrated solution is frequently used as the electrolyte in alkaline electrolyzers to transfer OH⁻ anions. The electrodes and the generated gases are physically separated by a porous inorganic diaphragm or separator that is permeable to the KOH solution [77]. Conversely, PEM, AEM, and SOEC separate the electrodes using an electron-insulating solid electrolyte. The ion transport occurs within the PEM, AEM, or solid oxide module. Thus, these specific electrolyzers eliminate the requirement for a liquid electrolyte solution [78]. A summary of the essential elements and operating parameters for each of the four types of electrolyzers is provided in

Table 3. Four types of water electrolyzers [80].

	Alkaline	PEM	AEM	Solid Oxide
Operating temperature	70–90 °C	50–80 °C	40–60 °C	700–850 °C
Operating pressure	1–30 bar	<70 bar	<35 bar	1 bar
Electrolyte	Potassium hydroxide (KOH) 5–7 mol/L	PFSA membranes	DVB polymer support with KOH or NaHCO3 1 mol/L	Yttria-stabilized zirconia (YSZ)
Separator	ZrO2 stabilized with PPS mesh	Solid electrolyte (above)	Solid electrolyte (above)	Solid electrolyte (above)
Electrode/catalyst (oxygen side)	Nickel coated perforated stainless steel	Iridium oxide	High-surface-area nickel or NiFeCo alloys	Perovskite-type (e.g., LSCF, LSM)
Electrode/catalyst (hydrogen side)	Nickel coated perforated stainless steel	Platinum nanoparticles on carbon black	High surface area nickel	Ni/YSZ
Porous transport layer anode	Nickel mesh (not always present)	Platinum coated sintered porous titanium	Nickel foam	Coarse nickel mesh or foam
Porous transport layer cathode	Nickel mesh	Sintered porous titanium or carbon cloth	Nickel foam or carbon cloth	None
Bipolar plate anode	Nickel-coated stainless steel	Platinum coated titanium	Nickel-coated stainless steel	None
Bipolar plate cathode	Nickel-coated stainless steel	Gold-coated titanium	Nickel-coated stainless steel	Cobalt-coated stainless steel
Frames and sealing	PSU, PTFE, EPDM	PTFE, PSU, ETFE	PTEF, silicon	Ceramic glass

. Conditions and components that show a great deal of variation amongst different producers or research and development organizations are indicated by highlighted cells. Additionally, this presentation sheds light on less developed technologies, as seen in the solid oxide and AEM categories in particular [79].

7. Requirements for Efficient Working

The substance which is employed as a photocatalyst must meet a number of functional requirements in terms of electrochemical energy and bandgap properties. These include a higher charge segregation rate, lower charge recombination, and higher charge transference and utilization. Moreover, higher chemical stability against photocorrosion and corrosion in aqueous environments are the prerequisites for economical water splitting. Therefore, to create materials having activities in water splitting during irradiation with visible light, particular photocatalytic surface properties and bulk properties along with energy structures are required [4].

The bottom of the photocatalyst’s conduction band (CB) should be at a more negative potential than the proton reduction potential (H⁺/H₂ = 0 V vs. the normal hydrogen electrode (NHE) at pH 0) in order to attain total water splitting. Additionally, in order to promote water oxidation, the photocatalyst’s valence band edge needs to be higher than that of the oxidation potential of water (+1.23 V vs. NHE at pH 0) [81]. Thus, based on these figures, the total energy needed to carry out the water-splitting reaction is 1.23 eV for a theoretical semiconductor’s bandgap. Theoretically, 70% of all solar photons are accessible for water splitting (

Table 4. Distribution of energy in the terrestrial solar spectrum (AM 1.5) [82].

Spectral Region	Wavelength [nm]	Energy [eV]	Contribution to Total Spectrum [%]
Near-UV	315–400	3.93–3.09	2.9
Blue	400–510	3.09–2.42	14.6
Green/yellow	510–610	2.42–2.03	16.0
Red	610–700	2.03–1.77	13.8
Near-IR	700–920	1.77–1.34	23.5
Infrared	920–1400	1.34- 0.88	29.4

) because that energy is equal to the energy of a photon with a wavelength of about 1010 nm.

There are some requirements for photocatalyst efficiency: First, it must be a semiconductor, meaning that it must have a bandgap. Secondly, the bandgap value should be around 2.0 eV to enable the catalyst to utilize a significant proportion of visible as well as UV light. Thirdly, band edge positions should be optimized to produce holes and excited electrons at the proper energy levels, where hydroxyl group anions and hydrogen cations are reducing and oxidizing, respectively. Fourth, there must be no imperfections in the structure of the photocatalyst in order to minimize the electron–holes recombination [83].

The band positions exactly meet the thermodynamic conditions needed to split water [84]. The next three factors appear to be crucial: (i) a molar ratio of 2, (ii) a turnover number (TON) that is catalytically acceptable, and (iii) a direct measurement of molecular oxygen and hydrogen that does not rely on current flow. These would represent the bare minimum needed for investigations on water splitting [85]. The production and separation of efficient light-induced carriers, as well as the adsorption of molecules of water at the catalytic center before water-splitting redox reactions, are the two main steps in an efficient photocatalytic water-splitting reaction. A directly Z-scheme heterostructure may be capable of achieving the exacting standards needed for photocatalytic water splitting [86].

8. Various Catalysts for Water Splitting

Designing a catalyst which could grab maximum solar light is the key task for enhanced photocatalytic water splitting. Excitation, bulk diffusion, and surface transmission are the three main mechanisms involved in photocatalysis. As a result, in order to be effective, a photocatalyst must fulfill several primary requirements regarding its chemical, crystallographic, and surface qualities. However, there are a few limitations in the semiconductors which must be addressed, and it is exceedingly challenging to locate a single element that can meet all of these needs. Thus, even though numerous semiconductors for water splitting have been created over the past few decades, most of them need sacrificial agents to yield hydrogen from water and work only in the UV region of the solar spectrum [87]. Numerous photocatalysts have already been designed and evaluated in recent years for both thermodynamically uphill reactions and downhill reactions, but TiO₂ continues to be the most extensively studied one due to its favorable physical and chemical properties. Additionally, conducting the process in the presence of organic compounds that act as a hole scavenger stimulates an irreversible reaction. This process enhances the kinetics of the water splitting to yield a higher concentration of hydrogen [88].

For photocatalytic water splitting, many catalysts are used, each with benefits and drawbacks. Some of them perform better than others in terms of efficiency. Co-catalysts are used in photocatalytic water splitting. Catalysts are elements or compounds that speed up or improve chemical reactions. The following

Table 5. List of co-catalysts used in photocatalytic water-splitting experiments [89].

Co-Catalysts	Examples
Noble metals	Au, Pt, Pd, Ru, and Ag
Transition metals	Ni, Cu, and Co
Metal oxides	CuO, NiO, and Cu2O
Metal sulfides	NiS, CuS, MoS2, and WS2

shows some co-catalysts used in photocatalytic water splitting.

9. Titania for Water Splitting

In the 1970s, it was found that TiO₂ reveals a lot of ideal assets which include its chemical stability, nontoxicity, and cost-effectiveness, as well as its Earth abundance. All these traits make it a suitable and ideal candidate for water splitting to generate hydrogen [90]. However, TiO₂ also has a broad bandgap (3.0–3.2 eV), which lowers the possibility of visible-light absorption [91]. By increasing the number of active sites and enhancing the electrical conductivity, the structural and chemical properties of TiO₂ allow for the engineering of various parameters such as the bandgap, light-absorption properties, and recombination time [92]. There are multiple polymorphs of TiO₂ with varying behaviors. As seen in Figure 15, the most prevalent ones are rutile, brookite, and anatase. The most common polymorphs for photocatalytic water splitting are rutile and anatase TiO₂; however, some attempts have also been made using amorphous TiO₂ (aTiO₂) as shown in Figure 16.

Numerous endeavors have been undertaken to incorporate dopants in order to enhance the optical absorption of TiO₂. For instance, compared to pure anatase TiO₂, Zhang et al. discovered that 12.5% copper (Cu)-doped anatase TiO₂ displayed a wider absorption [95]. According to a theoretical analysis, the enhanced optical absorption is explained by the introduction of an unoccupied impurity continuum band by Cu at the top of the valence band. According to another theoretical analysis performed by Morgade and Cabeza in 2017, it was discovered that co-doping TiO₂ with C, N and Pt, V narrows the bandgap and beneficially impacts the position of the conduction and valence band edges [96].

Our review provides a great insight about the current technologies used for water splitting in the form of electrolyzers and gives an overview of the TiO₂-based catalysts used for water splitting. Additionally, it demonstrates various characterization techniques used for analyzing the charge dynamics of the photocatalyst which can be helpful in designing a suitable catalyst for water splitting. Moreover, it highlights various modification techniques used to improve the catalytic performance of TiO₂ for enhancing its water-splitting aptitude to generate a higher hydrogen yield.

9.1. Photocatalytic Water-Splitting Mechanisms for the Production of Hydrogen Using TiO₂

The configuration of the electrical semiconductor plays an important role in the photocatalysis process. The two bands that separate semiconductors from conductors are the conduction band (CB) and the valence band (VB). The energy difference between these two bands is known as the bandgap (Eg). The valence band would include holes and electrons if not for the activation process. The energy level of incident light must match or exceed the energy level of the bandgap of the semiconductor in order to excite the semiconductor. When photons excite electrons in the valence band of a photocatalyst to the conduction band, they leave the holes behind and form pairs of electrons and holes [97]. The water-splitting photocatalytic process of TiO₂ may involve the following phases. First, as Equation (17) illustrates, when the surface is exposed to visible light, electron (e⁻) and hole (h⁺) pairs are formed. When the photogenerated electrons and holes quickly join in bulk or over the semiconductor surface, the energy is released as heat or photons. Electrons and holes can mix in the way shown by Equation (18), in addition to participating in chemical reactions. Moreover, as shown in Equation (19), the photoexcited carriers are capable of surface chemical reactions, in which the holes interact with H₂O molecules to produce H⁺ and •OH without recombining. Equations (20)–(22) ultimately demonstrate how photogenerated electrons and holes oxidize •OH and reduce H⁺ to generate O₂ and H₂ [97,98]. A simplified representation of the photocatalytic water-splitting mechanism is shown in Figure 17.

TiO₂ → TiO₂ (e⁻ + h⁺)

(17)

TiO₂ (e⁻ + h⁺) → TiO₂

(18)

h⁺ + H₂O → H⁺ + •OH

(19)

2e⁻ + 2H⁺ → H₂

(20)

OH + •OH → H₂O + 1/2O₂

(21)

H₂O_(l) → H_2(g) + 1/2O_2(g), (Overall reaction)

(22)

A significant positive change in the Gibbs free energy (G_o = +238 KJ/mol) indicates the thermodynamic uphill reaction of water splitting into H₂ and O₂. It has been determined that the band structure is a thermodynamic need for water splitting. Thus, at λ > 400 nm, the photocatalyst’s bandgap energy (Eg) needs to be higher than 1.23 eV. The photocatalyst bandgap should be greater than the reduction and oxidation potential of water. The bottom of the conduction band must be more negative than the H⁺/H₂ (0 V against the NHE), while the top of the valence band must be more positive than the O₂/H₂O oxidation and the bandgap of the conduction band must be compatible; thus, photoexcited holes and electrons can effectively oxidize potential (1.23 V). Therefore, 1.23–3 eV is the required bandgap range for an active semiconductor that produces hydrogen efficiently when irradiated with the visible light [100]. The development of a highly efficient semiconductor for this purpose is significantly influenced by two features of the processes and mechanism of photocatalytic water splitting using visible light: (i) The photocatalyst’s band structure need to be tuned for visible-light harvesting. (ii) The bulk or surface electron–hole recombination will not occur if the photo-induced charges in photocatalyst are successfully separated.

Numerous studies explored the idea of producing H₂ from water splitting using TiO₂-based photocatalysts. For example, a Pt/HS-TiO₂ photocatalyst, introduced by Zhu et al. [98], follows a similar mechanism of H₂ production. It has been discovered that H₂ evolution is significantly enhanced by the addition of a small amount of methanol as a sacrificial reagent. Acetic acid is formed when methanol is subjected to hydroxyl radicals. Water is converted into H⁺ and •OH more quickly when hydroxyl radicals are consumed, and electron–hole recombination is prevented on the surface of Pt/HS-TiO₂ by trapping electrons [97]. Moreover, the rate of reaction is increased when every available hole is completely utilized. In addition, water plays a crucial role in the electron–hole separation process by acting as a “sacrificial hole scavenger”. Hydrogen is produced via the reduction of H⁺, which is catalyzed by Pt nanoparticles on the photocatalyst surface [101].

9.2. Limitations of Solar Hydrogen Generation with TiO₂

It is quite challenging to produce an effective visible-light-active photocatalyst for commercial solar H₂ production to meet the energy requirements of industrial circular systems. The efficiency of TiO₂-based photocatalytic water-splitting systems, which produce hydrogen from solar energy, is currently relatively low. One of the most important reasons that TiO₂ nanoparticles have proven so successful in photocatalytic H₂ production is their role in charge separation. It has been noted that the excited electrons and hole pairs need to be moved and separated quickly in order to start the redox reaction over the photocatalyst’s surface. Charge carriers will reassemble on the surface of the TiO₂ photocatalyst or in bulk if the process proceeds slowly. On the TiO₂ surface, these sluggish processes lead to charge carrier recombination, which releases energy as heat or light. Moreover, as water breaks into H₂ and O₂ and releases energy, the reverse reaction recombining the elements back into water may occur with little difficulty. To enhance H₂ and O₂ separation and manage the reverse reaction, a variety of techniques can be used. Since the bandgap of different types of TiO₂ crystals ranges from 2.9 to 3.2 eV, UV light is the only light that can be utilized to generate H₂ in TiO_2. It is important to note that other types of materials having different bandgaps might also be appropriate for the production of H₂. The efficiency of the sun’s photocatalytic hydrogen generation has the potential to be boosted because visible light provides around 50% of all energy originating from the sun, while ultraviolet (UV) light only contributes 4% [102]. Researchers are working hard to enhance photocatalytic performance and increase visible-light sensitivity in order to make solar photocatalytic H₂ production a reality [103].

10. Strategies to Improve the Efficiency of the TiO₂ Photocatalyst

10.1. Doping

The contribution of the catalyst to improve water splitting can be seen by comparing the efficiency of hydrogen generation with and without the catalyst. In photocatalytic processes, carrier transferring and carrier trapping are both important. Therefore, doping TiO₂’s surface with metals or non-metals is a great way to boost the material’s electron–hole transmission [104]. Significant doping makes it more difficult to transport electrons and holes to the interface; as a result, metal ions are more likely to serve as recombination centers. Doping can be divided primarily into two types: (i) cationic doping and (ii) anionic doping. Both categories of doping have greatly altered TiO₂. New energy levels are created in TiO₂ above and below the valence band as a result of the anionic and cationic doping [105], while the latter frequently aids in altering the VB positions, reduces defects, and improves the chemical strength of TiO_2. The anionic doping is ascertained to minimize the bandgap energy and charge separation and promote visible-light absorption in TiO₂. The majority of the anionic doping in TiO₂ has been performed with N, C, S, and P. Due to the system’s higher stability, the N doping among them demonstrated considerably increased photocatalytic activity [106]. Similar to this, numerous elements were studied for their photocatalytic properties under UV–visible light and doped at the cationic site of TiO₂.

10.1.1. Anionic Doping in TiO₂

The characteristics of C, N, and S-doped TiO₂ nanomaterials were evaluated for visible-light absorption. It was found that TiO₂-P25 exhibits the conventional band-edge absorption at about 390 nm with the 3.2 eV bandgap, as well as the same values for C and S doping being observed. However, N-doping exhibited an absorption at around 415 nm with a bandgap energy of 3.0 eV. Additionally, an intriguing aspect of their valence band X-ray photoelectrons’ spectra showed that the addition of C, S, and N to TiO₂ resulted in the creation of new states (Figure 18). It was discovered that these extra states added deeper stages into the bandgap of TiO₂ in the sequence of C > N > S, and they were related to the C 2p, S 3p, and N 2p orbitals [107].

N-Doped TiO₂

Recent studies have found that oxygen vacancies lead directly to donor states underneath the conduction bands which could enhance the visible-light photoactivity, with substitutional nitrogen which inhibits the electron–hole pair recombination. It was hypothesized recently that the bandgap of nitrogen-doped TiO₂ was reduced below the inherent bandgap edges and that the photoactivity beneath visible light was caused by substitutional nitrogen atoms. It was determined that the nitrogen-induced effect that permitted sub-bandgap excitation was caused by a single N (2p) state rather than compressing the bandgap. Optical absorption revealed the localized N states in the TiO₂ lattice and substitutional doping, as well as the status of N as an anion-like (N-) ion with N-Ti-O as its chemical environment.

While oxygen vacancies were mainly responsible for absorption above 500 nm, and for absorption below 500 nm, the nitrogen states above the valence bands were primarily responsible [108].

F-Doped TiO₂

It was discovered that F-doped TiO₂ has a great photocatalytic efficiency for the breakdown of acetaldehyde gas when exposed to visible light and UV. The researchers came to the conclusion that F-positive doping’s effects, including the generation of oxygen vacancies, increased surface acidity, and also an increase in active sites, were responsible for the high photocatalytic activity. TiO₂ thin films with carbon doping exhibited hydrophilicity when exposed to visible light. The localized C (2p) generated above the VB is thought to be the cause of the material’s reduced sensitivity to visible light, which has a lower hydrophilic coefficient than UV light.

In order to maximize the photo-cleavage of water by visible-light irradiation, it was reported that the production of a vertical grown carbon-doped TiO₂ nanotube array offers high aspect ratios. Under visible-light irradiation (>420 nm), the synthesized TiO₂-C_x nanotube arrays demonstrated significantly greater densities of current and more effective water splitting than pure TiO₂ nanotubes. Under white-light illumination, the overall photo current was 20 times more as compared to a P-25 nanoparticle film [108].

S-Doped TiO₂

S-doped TiO₂ photocatalysts were made using a template-free oxidant peroxide approach and then crystallized using a hydrothermal method. It was determined that adding S to TiO₂ will replace oxygen atoms with sulphur anions, causing the bandgap to narrow as shown in Figure 18. The visual absorption of TiO₂ from UV to the visible area was significantly increased as a result. It was discovered that adding electron-deficient S atoms to the TiO₂ crystal lattice successfully trapped photo-induced electrons and decreased charge carrier recombination. Additionally, surface oxygen vacancies created by S-doping served as a location for the photogenerated electrons to be trapped, decreasing the recombination of electron–hole pairs.

The increased visible-light photocatalytic activity of S-doped TiO₂ for the degradation of methyl orange (MO) was the consequence of the harmonious effect of the effectively reduced charge carrier recombination and prolonged visible-light absorption [109].

10.1.2. Cationic Doping

In order to increase TiO₂’s photocatalytic activity, a great deal of research has been performed recently on both doping the material with transitional metal ions and doping it with rare-earth metal ions. It contributes to the creation of acceptor energy levels underneath the conducting bands of the material or donor energy levels above the valence band of TiO₂. Choi et al. [110] carried out a detailed study of the photoreactivity of metal ions doped into TiO₂ and their results were reported. It was found that doping the material with metal ions could shift the photoresponse of TiO₂ further into the visible spectrum. In a similar manner, Zhao et al. [111] synthesized bismuth (Bi)-doped TiO₂ nanotubes that were created by Zhao et al. [65] and hydrothermally treated in a 10-mol·L⁻¹ NaOH solution. They noticed that altering the Bi/Ti ratios would produce comparable outcomes. Because of their enhanced photocatalytic efficiency and increased ability to absorb light into the visible spectrum, the produced catalysts were able to produce hydrogen from a mixed glycerol/water solution more effectively than pure TiO₂ nanotubes. The Bi-doped TiO₂ catalyst would change the energy level of Bi ions and produce new Bi (3+x⁺) species. Resultantly, this leads to a decrease in the rate of electron–hole pair recombination. Noble metals, such as gold, silver, ruthenium, and platinum, have been found to be superior to other metals in the creation of H₂ because of their exceptional ability to increase photocatalytic performance into the visible spectrum by using strong surface plasmon bands [83,112]. For instance, Ortiz et al. [67] created TiO₂–F, an undoped TiO₂, and TiO₂Ag–F, a doped TiO₂, using the sol-gel/solvothermal method (SGH). TiO₂Ag–F was shown to exhibit an outstanding catalytic activity, producing 180 μmol of H₂ per gram of catalyst during a 4 h irradiation period. One important element influencing how quickly these materials exhibit photocatalytic activity when electrons are brought to them is the impossibility of electron–hole pair recombination, which is required to create metallic Ag clusters in TiO₂.

Furthermore, Ni et al. illustrated how different transition metal ions alter photocatalyst activity via diverse electron–hole migration pathways in order to discuss the doping of TiO₂ with transition metals, comparable to that of noble metals, as shown in Figure 19 [97]. Because transition metal nitride materials have virtually comparable optical properties to noble metal nitride materials, they have lately found use as plasmonic materials for photocatalytic processes [113]. Transition metals transport more electrons into the TiO₂ conduction band because they are more chemically stable and have a higher plasmon resonance than noble metals [114]. Consequently, the photoelectron transfer from the transition metal to the CB of TiO₂ results in an efficient reduction process at the semiconductor’s surface. For example, to examine the photo-thermochemical cycle (PTC) mechanism, Xu et al. [115] produced films of TiO₂ doped with iron (Fe) using a sol-gel process. Fe-doped TiO₂ outperformed untreated TiO₂ in the PTC mechanism’s water-splitting method for creating H₂.

As previously mentioned, cationic doping effectively establishes intra-band energy levels near to the conduction band of TiO₂, which causes a red shift in the system’s optical properties [116]. This effect has also been reported in cations such as transition metals, rare-earth metals, or other metals doped with TiO_2. The main disadvantage of cation doping seems to be the increased number of charge carrier trapping sites (both for electrons and holes), which significantly lowers the photocatalytic efficiency.

This is due to the fact that the trapped carriers often merge back with their corresponding mobile carriers. As shown in Figure 20, the process of cationic doping fundamentally involves tuning both of the electronic structure and the Fermi level of the d-electron configurations in TiO₂ in order to tailor the levels of energy for the absorption of light in the visible region.

Under visible-light irradiation, iron was utilized to dope TiO₂, and its photocatalytic activity outperformed that of the Degussa P-25 commercial catalyst. Fe-doped TiO₂ was discovered to have a far higher photocatalytic efficiency than Degussa P-25 when exposed to visible light. In the TiO₂ lattice, Fe³⁺ cations served as shallow traps. After doping at a comparatively low level, the best photocatalytic characteristics were attained. The kinetics of the recombination process, which was correlated with spacing among dopant cations within the TiO₂ lattice, was closely tied to this. In addition to trapping electrons and holes, Fe ions significantly increase photoactivity. Due to a reduction in the surface-active centers’ density, the highest photoactivity was seen at 0.5 weight percent of the Fe level. It was discovered that the additional Fe atoms soluble in the TiO₂ phase had a rutile micro-structure with an average grain size of less than 10 nm. Compared to conventional P-25 powder, Fe-doped powder has a greater absorption threshold in the 427–496 nm range (406 nm). Fe-doped powder’s color turns from white to brilliant yellow, and the bright yellow becomes dark yellow when the Fe level is over 4.57 weight percent [106].

It is important to remember that optimal performance requires the precise management of doping levels or concentrations. If the dopant concentration is extremely high and functions as a recombination center, the photocatalytic performance can be compromised. Similar to the study of Zhao et al. [110], the transformation of anatase-phase nanotubes into titanate-phase nanobelts results from an excess of Bi ion doping in the TiO₂ crystal structure. It is anticipated that this alteration will lower the catalytic efficiency in terms of hydrogen yield. Therefore, in order to optimize and improve the efficiency of charge trapping and separation, it becomes crucial to accurately control the dopant quantity [91].

10.2. Sensitization of TiO₂

The excitation of a chromophore, like an organic dye, is a useful technique for filling the conduction band of a large bandgap semiconductor with electrons under visible light [117].

The two types of sensitization are composite semiconductors and dye sensitization. To use visible light for textile waste water energy conversion, dye sensitization is frequently used. The table below lists common dyes. The dyes’ wavelengths fall between 442 and 665 nm. For solar cells and photovoltaics, some dyes with redox properties and visible sensitivity to light are used. When compared to bulk TiO₂ photocatalysts, highly distributed TiO₂ species produced in zeolite frameworks, SiO₂, B₂O₃, and Al₂O₃ matrices demonstrated a superior and distinctive photocatalytic performance [107].

Table 6. Some commonly used dyes and their classifications [107].

Class	Dye
Thiazines	Thionine, methylene blue, new methylene blue, azure A, azure B, azure C
Hiazines	Toluidine blue
Phenazines	Phenosafranin, safranin-O, safranin-T, neutral red
Xanthenes	Fluorescein, erythrosin, erythrosin B, rhodamin B, rose Bengal, pyronine Y, eosin, rhodamine 6G
Acridines	Acridine orange, proflavine, acridine yellow
Triphenyl methane derivatives	Fusion, crystal violet, malachite green, methyl violet

shows some commonly used dyes and their classifications.

10.2.1. Dye Sensitization

In this method, a narrow bandgap semiconductor is coupled with a higher negative conduction-band level with a wide-band semiconductor having a larger bandgap. This increases the mixed photocatalyst’s ability to absorb light by injecting conduction-band electrons into the larger bandgap from the small bandgap [72]. In this method, solar photons are used to excite the dye molecules attached to TiO₂ in order to produce highly excited electrons, which are subsequently transferred towards the CB of the TiO₂. The TiO₂ conduction band’s electrons are then transferred to interfacial electron acceptors (such as H₂O, I³⁻, H⁺, and O₂) and cause a number of redox processes, as shown in Figure 21. In order to reach a high degree of photoconversion efficiency, interfacial energy transfer from the excited dye to the TiO₂ conduction band and effective solar energy usage should be the critical characteristics.

The production of organic dye-sensitive TiO₂ particles for the oxidation of 4, chlorophenol and Cr(VI)) pollutants under visible light (>420 nm) was described, and the impact of different anchoring groups was examined. To achieve this, they created Ru-free organic dyes in a donor–acceptor configuration with various numbers (n = 1, 2, and 3), known as D1, D2, and D3, respectively [108].

The existence of electron donors such tri-ethanol-amine (TEOA) as well as ethylene-diamine-tetra-acetic acid (EDTA) enabled researchers to discover that all three types of dyes are effective for producing hydrogen gas, with the ordering being D3-D2 > D1 (TEOA) and D3 > D2-D1 (EDTA). According to the results of the FTIR research, D1 and D3 are most likely anchored to the surface of TiO₂ using bidentate modes with single and double carboxylates, accordingly. They also looked into the several ways that D2 is anchored, based upon competing electron donors. Because of the various photochemical circumstances and reaction routes, they showed that the carboxylate number becomes less significant in the sensitized pollutant oxidation. Figure 22 shows the tri-branched organic dye molecules (sensitizers D1, D2, and D3, respectively) having mono-, di-, and tri-carboxylate attaching groups.

10.2.2. Sensitization with Noble Metal Particles

To sensitize TiO₂ with metallic nanoparticles is necessary by using the localized surface plasmon resonance (LSPR) phenomenon in order to increase the visible-light absorption because TiO₂ can only be active in the UV region. Au, Ag, Pt, and Pd, as well as their alloys, are the main noble metal nanoparticles due to their absorption in the visible light region [118]. The LSPR effect, which serves as an electron sink, can “excite” the electrons in the CB under visible-light irradiation to produce extremely energizing “hot electrons” [119].

The “hot electrons” then travel directly to the CB of TiO₂, creating a metal–semiconductor Schottky junction. The LSPR effect, meanwhile, may create a powerful local electronic field that increases its conduction electrons’ energy and makes it easier for them to transfer to and interact with electron acceptors. Resultantly, the electrons and holes produced by photons are effectively segregated [120]. It is possible to decorate TiO₂ restructurings with noble metal nanoparticles using hydrothermal, electrospinning plasma sputtering, UV reduction, electrode location, and other techniques. Hollow Pt/TiO₂ spheres have been created using a straightforward sol-gel process. Due to the increased donor density and decreased recombination of photogenerated electron–hole pairs, Pt/TiO₂ hollow spheres exhibit a greater photocatalytic water-splitting activity than pure hollow TiO₂ spheres under visible light [121]. Photocatalytic splitting by Ag/TiO₂ based on the LSPR effect is shown in Figure 23.

10.3. Heterojunction Construction

One of the most successful methods to address the aforementioned problems is the creation of a heterojunction based on TiO₂ by adding an additional low-bandgap conductor [122]. In hybrid systems, there are two distinct techniques used: hetero- and homojunctions. The use of a heterojunction in design is a typical strategy because it has the ability to significantly reduce serious charge recombination and maximize the use of visible light. On the contrary, the production of a homojunction (phase-junction) between two separate phases might modify the transit of the charge carriers along the inter-phase band alignment, making it a different configuration access to charge separation [123]. The variation in the photo-induced electron–hole transfer mechanism, which may be split into two types as follows, determines the heterostructure of the photocatalyst. The mechanisms are (i) the Z-schemes charge transfer mechanism, as well as (ii) the Type II heterojunctions transfer method [124]. The photocatalytic performance of a carefully designed heterojunction can remarkably enhance the efficiency of the hybrid photocatalyst as compared to the unassisted photocatalyst, according to the literature. More significantly, it has been discovered that a hybrid photocatalyst performs better for a composite heterojunction photocatalyst when two photocatalysts are mixed in a ratio of 1:1 [125].

There are three known forms of band alignment in heterojunctions as depicted in Figure 24 [87]. The valence band of SC-1 is more positive than that of SC-2 while the conduction band of SC-1 is more negative than that of SC-2 in Type I. In Type II, the valence band of SC-1 is less positive than that of SC-2, while its conduction band is more negative than that of SC-2. In Type III, both the valence band and conduction band of SC-2 are lower than those of SC-1.

In a semiconductor–semiconductor heterojunction, holes and electrons are thought to move to the less positive valence band and the less negative conduction band, respectively. In this case, SC-2 during Type I alignment, holes and electrons end up accumulating in the semiconductor with the smaller bandgap. Therefore, the chances of their recombination increase, preventing any enhancement in photocatalytic activity. Improved charge carrier separation will be attained in Type II alignment, because the electrons and holes in the two semiconductors will be separated. On the other hand, in Type III alignment, neither electrons and nor holes will be able to transport to one another. They are therefore considered to be two separate semiconductors and do not form a heterojunction. A semiconductor’s Fermi level is the total chemical potential of its electrons, and it is typical for different semiconductors to have unequal Fermi levels, so a semiconductor–semiconductor heterojunction is expected to form an electric field similar to that of a Schottky junction, where the flow of charge carriers is facilitated in one direction but hindered in another direction.

For the Type I alignment, the charge separation will not be encouraged as explained above. It is possible that just a very small internal electric field can arise because both semiconductors are intrinsic. Better charge separation is envisaged for a heterojunction with a non-negligible internal electric field. In the case where SC-1 has a higher Fermi level than SC-2, free electrons from SC-1 will flow to SC-2 during the internal electric field formation until the Fermi levels are aligned (Figure 24, Type I), resulting in the formation of a space charge region close to the interface. The internal electric field flows from SC-1 to SC-2 because of the electrostatic interaction that causes the surface of SC-2 to be negatively charged and the surface of SC-1 to be positively charged [127]. Thus, after photoexcitation, the internal electric field promotes the transfer of holes from SC-1 to SC-2 while blocking the movement of electrons in the same direction. Instead, because of the more negative conduction band of SC-1, the electrons from SC-2 will move towards SC-1 but gather at the junction [128]. In a heterojunction between an n-type SC-2 and a p-type SC-1, where the conditions are reversed (Type I-2 in Figure 25), electrons are more likely to move from SC-1 to SC-2, whereas holes from SC-2 gather at the interface. In both scenarios, it is theoretically possible to attain an improved charge separation.

For the Type II band alignment, when the Fermi level of SC-1 is greater than that of SC-2, an internal electric field is created that points from SC-1 to SC-2 (Type II-1 in Figure 25). In the space charge region for SC-1, photogenerated electrons will flow towards interior, while holes will move in the direction of interface; in the same way, the photogenerated charge carriers for SC-2 will flow in the opposite direction as for SC-1. Thus, the interface intercepts both the electrons from SC-2 and the holes from SC-1, which leads to their annihilation via recombination [129]. For catalytic reactions, however, the electrons in SC-1 and holes in SC-2 are retained without losing their reduction and oxidation ability, respectively. This kind of heterojunction is also known as the Z-scheme system because the photoexcitation and charge recombination processes show a Z-shape route in the corresponding energy diagram [130]. The internal electric field flows from SC-2 to SC-1 (Type II-2 in Figure 25), when the Fermi level of SC-2 is greater than that of SC-1 in the Type II band alignment, such as in the scenario where SC-1 is a p-type semiconductor and SC-2 is an n-type. The electrons of SC-1 will move in the direction of the interface. The electrons of SC-1 can cross the interface and move to the conduction band of SC-2 because the conduction band of SC-2 is less negative than the conduction band of SC-1. In a similar manner, the holes of SC-2 will move to the valence band of SC-1.

For the Type III-1 heterojunction, no electrons and holes can cross the junction and enter the corresponding bands of another semiconductor (Figure 25). The energy level barriers prevent the transport at the interface even though both the electrons and holes of SC-1 and SC-2 will move towards the junction. A twisted Z-scheme system can be created by installing an appropriate charge mediator to act as a bridge connecting these two semiconductors [87]. The leftover electrons and holes in a twisted Z-scheme system will ultimately wind up in the more negative conduction band and more positive valence band of the two semiconductors, respectively, similar to what a conventional Z-scheme system can achieve.

TiO₂-based heterojunction semiconductors have been the subject of several studies. One such example is the solvothermal fabrication of an anatase TiO₂/g-C₃N₄ heterojunction doped with boron. The heterojunction’s enhanced light-absorption efficiency and improved charge separation resulted in an amazing visible-light photocatalytic performance (47.3 mol h⁻¹) [131]. Similarly, under solar light irradiation. Liu et al. investigated the Ag₃PO₄-TiO₂ heterojunction’s band structures and its effectiveness in photocatalytic water splitting for hydrogen evolution. The determination of the Fermi levels of TiO₂ and Ag₃PO₄, i.e., −5.09 eV and −5.95 eV, respectively, led to the evaluation of their energy band diagrams. It was revealed that the heterojunction formed with 12 wt% Ag₃PO₄ had the fastest rate of hydrogen evolution (44.5 μmol g⁻¹h⁻¹), which was 5.7 times faster than TiO₂. When the heterojunction was deposited with Au nanoparticles, hydrogen generation enhanced 10.2 times (453.0 mol g⁻¹h⁻¹). It was discovered that the hydrogen evolution performance was consistent with the PL analysis results.

He et al. [132] successfully manipulated the nucleation and growth processes to form an edge-on heterojunction of MoS₂/TiO₂. This heterostructure may enhance interfacial electro-conductivity and alter the electron transfer pathway along basal planes, thereby improving the separation of photogenerated charges and, ultimately, enhancing the catalytic abilities of TiO₂. This study mainly focused on the simple insertion of MoS₂ into photocatalysts to generate heterojunctions, with minimal focus on the systematic design and management of nanoscale architectures to maximize light absorption, charge separation, and surface catalytic reactions. To achieve a higher photocatalytic efficiency, Sun et al. [133] developed a particulate hierarchical hollow tandem heterojunction photocatalyst to improve light utilization in a broad solar spectral range. It comprised two heterojunctions of MoS₂/CdS NPs and hollow black TiO₂/MoS₂ nanosheets. To prevent photogenerated electron and hole recombination, the MoS₂ nanosheets in this instance served as both a co-catalyst and a bridge to integrate hollow black TiO₂ and CdS into a particulate tandem system (referred to as b-TiO₂/MoS₂/CdS). This architecture is different from the conventional single heterojunction architecture and similar to a Z-scheme structure. In order to harvest UV–Vis, Wang et al. [134] also developed a Z-scheme Cu₃P/TiO₂ photocatalyst that does not require a noble metal. The optimized Cu₃P/TiO₂ photocatalyst performs significantly better than TiO₂ under solar light irradiation, with a hydrogen evolution rate of 607.5 mol h⁻¹g⁻¹, which is 30.4 times greater than the bare TiO₂ photocatalyst. Cu₃P/TiO₂ exhibits enhanced charge carrier separation efficiency, as demonstrated by the PL spectra, transient photocurrent response, and electrochemical impedance spectroscopy (EIS). In this study, photoexcited electrons are enriched on Cu₃P nanosheets for H₂ evolution, whereas the holes on the TiO₂ are left.

As discussed above, every kind of heterojunction has unique charge transfer properties. As the heterojunction and band alignment type are tightly associated, understanding the latter is crucial to understanding how charge carriers flow within heterojunctions. Assigning the band edge positions in the energy diagram makes it simple to draw the band alignment; therefore, understanding the band edge positions of the component semiconductors in a particular heterojunction is crucial in this regard. Additionally, band edge positions show the heterojunction’s redox potential; a stronger reduction strength is indicated by a more negative conduction band, while a stronger oxidation ability is indicated by a more positive valence band.

10.3.1. Type II Heterojunction

The Type II heterojunction is the one that has the most researchers’ interest out of all three typical heterojunctions. The level of the CB and VB of semiconductor-II is higher than that of semiconductor-I in Type II heterojunctions [135] as shown in Figure 26. Additionally, the band bending phenomena caused by the disparity in chemical potentials can be used to stimulate the passage of charge carriers in the opposing directions. The Type III heterojunctions’ band structure (Figure 12) is comparable to that of the Type II heterojunctions, with the exception that the staggered gap widens to the point where the bandgaps do not overlap [136].

TiO₂ can be used in heterojunctions with some other semiconductors to boost redox reactions and slow down quick charge recombination, which enhances H₂ production [57]. TiO₂-based materials often combine the lower CB of photosystem II with the high CB of photosystem I (PS I) to generate conventional heterojunctions (PS II). The photogenerated holes on that VB of PS I will remain after the photogenerated electrons of PS I are transported to PS II under light irradiation. The recombination of the charge carrier produced by the light can be prevented in this fashion, improving the efficiency of the photocatalytic reaction. For various semiconductors, the reduction and oxidation reactions are simultaneously taking place. The redox reaction, however, takes place at a lower redox potential but also forgoes its capacity to be maximized.

To separate the photogenerated electron–hole pair and therefore increase the activity of photocatalytic water splitting, the Type II heterojunction is crucial. In Type II heterojunctions, the inner host, such as TiO₂ (host) with CdS, has a wide bandgap whereas the sensitizers or outer shell typically has a small, mid bandgap and exhibits substantial visible-light absorption (sensitizers) [137]. TiO₂-based Type II heterojunctions come in two different types due to the staggered band structures. Due to the effective charge separation, many conventional Type II heterojunctions were created using electrospun TiO₂ NFs and a variety of semiconductors, such as metal oxides, metal sulfides, metal halides, and other materials. In Type II heterojunctions, photo-induced electrons flow from the more negative CB to a less negative CB while holes flow simultaneously in the opposite direction, leading to an all-around effective e⁻/h⁺ separation [137].

10.3.2. p-Type or n-Type

An efficient method to increase photocatalytic performance for the production of hydrogen is to create a heterojunction with other n-type/p-type semiconductors. Because when a heterojunction is created, the local electric field that results causes holes to travel in the reverse direction from photogenerated electrons, improving the separation of photon generated hole and electron pairs and enhancing photocatalytic activity [138]. It is important to choose semiconductors for the heterojunction such that they have various band edge potentials and conducting kinds. This design offers the system various benefits, including an increase in (i) charge separation, (ii) charge carrier life, (iii) recombination resistance, and (iv) interfacial charge transit to the adsorption [139]. Because of the internal electric field produced by the charge carrier diffusion between the p-type and n-type semiconductors in p-n heterojunctions, photogenerated electrons and holes are ultimately driven in a separate direction, which reduces their tendency to recombine and increases their lifetime [137].

Compared to Type II, p-p, n-n, and p-n heterojunctions are proposed to maintain effective EHPs (electron–hole pairs), and these heterojunctions are obtained by combining both h⁺ rich and e⁻ rich semiconductors [140]. By creating an electric field at the interface, the heterojunction above speeds up the separation of the charge carriers. The potential difference of other linked semiconductors should influence the generated EHP migration from the surface (interface) to the interior bulk of PS. In n-n type and p-p type heterojunctions, the work function (Φ) difference is insufficient to produce a sufficiently strong electric field to isolate EHP [141]. Consequently, the photogenerated e⁻ and h⁺ can be effectively isolated by a p-n photocatalytic heterojunction due to the internal electric field produced by the majority carriers moving in opposite directions [142]. Before sunlight, the predominant charge e⁻ from n-type semiconductors tends to move to p-type, while h⁺ migrates from p-type semiconductors to n-type until both Fermi levels (E_f) are in equilibrium [143]. As shown in Figure 27b, under light irradiation, the formed EHPs efficiently separated due to the combined effect of the photoactive semiconductor’s band alignment and the generated charge region close to the p-n interface. Furthermore, since the CB and VB of p-type semiconductors are often higher than those of n-type semiconductors, EHP tunneling from one PS to another one is thermodynamically possible [144].

It is reported that in practical applications, certain external factors and surface properties such as surface states, surface adsorption, and interactions between the e-donor and acceptor, may cause band bending near the semiconductor junction interface, taking into account the charge migration mechanism. Allocation takes place between the bulk and surface of both semiconductors in order to align the E_f charge. If there is an excess of surface e⁻, the bands will bend upward; if there is an accumulation of h⁺, the bands will bend downward. The separation and diffusion of produced charge carriers can be accelerated by band bending in semiconductor–semiconductor heterojunctions and the induced electric field resulting from charge redistribution [145]. In p-n junctions, the degree of band bending is somewhat controlled. Until their acceptor levels are dynamically in resonance with the Fermi level, the e⁻ acceptors receive charge from the semiconductor. A space charge region and related band bending are the outcome of the ground state charge transfer that occurs across the interface. The interaction between the substrate and the adsorbate is significantly weakened by this space charge area, which can also influence the charge transport characteristics across the interface. These parameters are difficult to regulate and not always well known in experiments.

Furthermore, the redox potential and surface Φ are used to determine the band bending close to the semiconductor surface. A photoactive semiconductor employed in the oxidative side often has a greater Φ than the reductive PS. When they come into contact with one another to create equilibrium, the semiconductor with the greater Φ causes downward bending towards the interface and vice versa [145]. The downward band bending transports the photoexcited e⁻ to the surface to carry out surface reduction reactions, whereas upward bending can move the h⁺ to speed up the oxidation reactions. The tunneling rate of photoexcited e⁺ and residual h⁺ in the p-n junction, as shown in Figure 27 is significantly affected by the direction of the internal electric field generated near the junction created by the contact between photoactive semiconductors with different Φ. Band bending may have an impact on two processes: (i) photocatalytic reactions by minimizing recombination and speeding up the tunneling rate of photoinduced EHP; and (ii) thermal absorption and also desorption of produced similar products. Numerous variables, including the kind of dopant employed, its concentration, particle size, surface phase, and chemistry affected the degree of band bending [146]. One charge carrier’s tunneling was increased by band bending to the surface of the semiconductor, either it is e⁻ or h⁺. This could enhance the associated e⁻- or h⁺-mediated photocatalytic reaction occurring on the semiconductor’s surface. To preserve the band bending and neutrality of the photocatalytic material, the inverse charge carriers may also tunnel across the internal electric field in the opposite direction and become depleted [147].

p-n junctions based on transition metal oxide semiconductors have been developed and designed because of their exceptional stability. Because of their robust e⁻-e⁻ interaction, configurational simplicity, and stoichiometric diversity, they provide an excellent broad range of photocatalytic applications [148]. Furthermore it has been observed that heterojunctions based on transition metal oxides exhibit excellent photostability and a highly effective recycling rate [149]. In order to get over these bottlenecks, a number of p-n transition metal oxide-based photocatalysts have been investigated.

A number of benefits are provided by efficient p-n type transition metal oxide-based heterojunctions, such as (i) more efficient charge carrier isolation; (ii) a rapid EHP tunneling rate; (iii) a longer photoexcited EHP lifespan; and (iv) the separation of locally inappropriate redox reactions. The p-n type heterojunctions based on transition metal oxides have a high photocatalytic output due to all of the previously described features [150].

There are several simple methods for fabricating p-n junctions based on transition metal oxides, all of which require readily available precursors. Choosing two photoactive semiconductor materials with h⁺ and e⁻ as the majority charge carriers is essential for effective synthesis procedures. Their distinct morphologies, band edge locations, Fermi level alignment, and interfacial characteristics all play an important role in the formation of an interfacial contact between the two photocatalysts. To achieve the optimum photocatalytic results, the features of metal oxide-based semiconductors should be completely utilized. Depending upon the morphological adjustments and requirements, their characteristics can vary greatly [151]. One or more steps may be involved in the fabrication strategies, depending on the specific requirements. p-n heterojunctions are created by adjusting a number of variables, including pH, temperature, surfactants, solvents, reaction time, and different kinds of transition metal oxides. A variety of techniques including solvothermal [152], sol-gel [153], self-assembly [154], hydrothermal [155], chemical-solution–deposition–decomposition [156], etc., are successfully applied to the construction of photoactive p-n heterojunctions based on transition metal oxides with appropriate morphology. The predicted characteristics and appearance of as-constructed p-n heterojunctions were analyzed using suitable characterization techniques in order to verify the effectiveness of the implemented fabrication routes.

The most facile and economical ways to modify the surface characteristics of metal oxides are by hydrothermal and sol-gel synthesis, among the other synthetic processes already described. The main benefit shared by both approaches is the achievement of stable surfaces and a large surface area. Therefore, the majority of publications have used sol-gel and hydrothermal techniques to create p-n junctions based on transition metal oxides. We examine the photodegradation ability of the NiO/TiO₂ p-n junction from several research papers in order to investigate the impact of fabrication procedures on photoactivity. The two published papers discussed the synthesis of p-n heterojunctions in response to a model methylene orange pollutant [157]. Lin et al., constructed a p-NiO/n-TiO₂ composite by employing the chemical-solution–deposition–decomposition technique and discovered a 92% photodegradation rate for methylene orange. In contrast, Jiang et al. [158] synthesized a NiO/TiO₂ composite using the sol-gel method, which remarkably displayed reduced photooxidation (68.3%, methylene orange) and higher photoreduction (87.0%). Thus, the development of the p-n heterojunction is entirely dependent on the employed manufacturing techniques, with the final objective being the coupling of p- and n-type semiconductor photocatalysts to maximize their respective properties.

Improved illumination absorption, photocorrosion stability, charge transfer efficiency, and ultimately photocatalytic efficiency are all provided by the construction of TiO₂-based p-n heterojunctions. This section contains information on various TiO₂-based p-n heterojunctions to enhance the properties of UV-active TiO₂ catalysts. Scientifically, under the influence of the internal electric field, excited e⁻ present in the n-type CB is directed to the VB of p-type semiconductors, promoting the EHP isolation rate. Because of its low sunlight-absorption range and extremely fast charge carrier assembly rate, onefold e⁻-rich TiO₂ is not usually a perfect semiconductor. Zeolites are effective materials with a large surface area, a low environmental impact, and unique photophysical characteristics that regulate the rate of EHP migration. Improved visible-light-responsive and extremely efficient CuO/TiO₂-zeolite (Na_1.89Al₂Si_2.88O_9.68) p-n hetero nanostructures were reported (Zhao et al.) using the conventional impregnation method, providing evidence for this claim. CuO is a significant transition metal oxide semiconductor with a p-type narrow bandgap with 1.2–1.5 eV that interacts with n-type wide bandgap TiO₂ with 3.2 eV. The CB, VB edge, and E_f of TiO₂ photocatalysts are higher than those of a CuO semiconductor prior to linking p-type CuO and n-type TiO₂ photocatalysts. During the coupling process, upshifting and downshifting of the E_f in CuO and TiO₂, respectively, was noticed, until the equilibrium stage was attained. At equilibrium, an effective p-n junction with a strong inner electric field was identified. This allowed photoexcited e⁻ from the p-CuO CB to n-TiO₂ and residual h⁺ from the n-TiO₂ VB to p-CuO, respectively, to migrate quickly. The remaining h⁺ on TiO₂ and isolated e⁻ on CuO were free to start photocatalytic redox reactions with the surface-adsorbed methylene blue (MB) dye. The photostability of an as-fabricated visible-light-responsive p-n CuO/TiO₂-zeolite heterostructure was demonstrated by its efficient reuse without any degradation in its photocatalytic function [159].

Similarly, mesoporous anatase TiO₂ hollow spheres having a low density, a large surface area, and crystalline nature showed exceptional optical output. For carrying out photoreactions, hybrid TiO₂ nanostructures demonstrated a configurable structure and also surface morphology. Wang et al., adopted the sol-gel technique to synthesize hollow TiO₂ coupled with NiO hybrid cells followed by a hydrothermal approach for H₂ evolution and RhB degradation. The usage of photoactive NiO (3.55 eV) and TiO₂ (3.3–3.2 eV) presents two main practical challenges: (i) wide bandgap energy, which demands high energy photons with wavelengths below 387 nm (UV area) and (ii) fast charge carrier recombination. It can successfully produce a p-n heterojunction with an n-type TiO₂ semiconductor, regardless of the NiO wide bandgap. As-fabricated hollow TiO₂ nanospheres had an absorption edge that was shifted from 397 to 553 nm, which increased the concentration of NiCl₂-6H₂O. As seen in Figure 28, anchored p-type NiO on the surface of n-type TiO₂ was stimulated by visible light to produce effective charge carriers after coupling. The inner electric field at the interface between n-type TiO₂ and p-type NiO was established through the migration of photoexcited e⁻ to the +ve side and photoremaining h⁺ to the –ve side. h⁺ in the VB of NiO interacted with H₂O to make highly reactive ∙OH radicals, while excited e⁻ in the CB of TiO₂ were trapped by O₂ to produce ∙O₂⁻. Highly fluorescent 2-hydroxyterephthalic acid was produced when terephthalic acid reacted with ∙OH, as demonstrated by the increased intensity to 425 nm in the photoluminescence (PL) spectrum, which verified the production of highly active ∙OH radicals [160].

Additionally, it is examined that connecting one TiO₂ molecule to another is a useful tactic for creating a nanohybrid with extensive properties. Hybrid TiO₂ structures were reported to achieve a 99.57% photodegradation efficiency with effective anodization followed by the impregnation approach in research [161]. As-synthesized n-type TiO₂ nanotubes (TONTs) were combined with another photoactive TiO₂ that had extra aluminum doping (overlayer) to create a h⁺-rich TiO₂ in order to take advantage of their structural benefits. Tiny p-n heterojunctions were created close to the TONT/Al: TONT interfaces when Al₂O₃ was deposited over the surface of TiO₂, creating a significant surface area for interaction. According to the X-ray photoelectron spectroscopy (XPS) spectra, doping of Al³⁺ ions in the lattice formed the acceptor level near the VB maxima and replaced some of the Ti⁴⁺ ions in the host. The semiconductors’ band structures, as shown in Figure 29, showed that the p-TONT’s CB and VB were both at higher energy levels than those of the n-Al: TONT, indicating that the EHP migration followed the Type II pathway rather than the Z-scheme [161]. Thus, it can be observed from the aforementioned publications that the development of TiO₂-based p-n junctions accelerated the development of their optical, physical, and chemical characteristics and significantly increased their total photocatalytic efficiency.

There are many techniques which offer the most important understanding of p-n junction faults or impurity (extrinsic) recombination states. This is a summary of some of the advanced characterization techniques, including cathodoluminescence (CL), transmission electron microscopy (TEM), electroluminescence (EL), scanning electron microscopy (SEM), electron beam-induced current (EBIC), deep-level transient spectroscopy (DLTS), scanning near field optical microscopy (SNOM), electron beam-induced current (EBIC), and secondary ion mass spectrometry (SIMS). Using electron beams, cathodoluminescence and electroluminescence are potent, costly, and labor-intensive methods for examining optoelectronic characteristics. Initially stimulating electron–hole pairs, the electron beam also aids in the local property detection of luminescence radiation. The matching defect structure distribution in the semiconductor junction is shown by the spatial distribution variation luminescence map. Defect kinds and various resistance distributions can be related to the luminescence emission wavelength.

When dealing with non-radioactive p-n junctions, the approach still loses its effectiveness. SEM or scanning tunneling microscopy (STEM) is used in conjunction with the electron beam-induced current mapping technique to examine the characteristics of defects and minority carrier diffusion lengths in p-n junction semiconductors. To measure the fluctuations in current across the p-n junction, electron beams are scanned over the entire sample; the defect states trap the electrons, lowering the current value. This technique works well for transitions that are radiative or non-radiative. Moreover, in order to convey different contrast mechanisms, NSOM uses optical excitation to scan the defect map signals transmitted through optical fibers.

Similar to this, DLTS is an extremely sensitive method that provides detailed information about electrically active flaws concealed within the forbidden gap. While the defect charge state retrieval roots the capacitance transient in accordance with its population density, the device scans at a constant capacitance value. TEM, a significant microscopic technology, was used to directly examine defect images of the p-n junctions with atomic-level resolution. Since TEM imaging uses a small quantity of the sample and does not properly validate the faults in the complete sample, its ability to confirm defects is indisputable. One sensitive surface analysis method is SIMS, which uses primary ions to sputter the sample layer. Additionally, elemental and crystalline defects are carried by secondary ion emission from the overlying surface, together with the diffusion-broadening characteristics of the p-n junctions.

The capacitance–voltage (C–V) method is another reported approach which is used for determining the breakdown voltage and doping density. However, it is discovered that this type of methodology makes data interpretation and analysis difficult. Above all, it can be challenging to calculate the accurate and real gadget properties. The p-n junction I-V method is a suitable approach that offers far more material information needed for the actual device construction and functioning. In both perfect and imperfect semiconductor crystal structures, recombination is a prevalent phenomenon. Excess recombination occurs in imperfect semiconductors because of flaws and impurities (traps) [106].

10.3.3. Z-Scheme Heterojunction

The direct Z-scheme photocatalyst idea was initially put forth in 2013 [163]. The band configuration and electron migration process of Z-scheme heterojunctions are shown in Figure 30. The electron transfer route among semiconductors is different in Z-scheme hetero junctions compared to Type II heterojunctions, although they have an identical band configuration. The path taken by electrons as they move between semiconductors resembles the letter “Z” in English [164]. The photogenerated electrons in semiconductor-II having a lower reduction ability mix with the photogenerated holes in semiconductor-I with a lower oxidation ability in the photocatalytic reaction. As a result, it is possible to preserve the photogenerated electrons in semiconductor-I having a high reduction ability and the photogenerated holes in semiconductor-II having a high oxidation ability.

Additionally, the photogenerated electron from semiconductor-II will migrate to semiconductor-I due to the electrostatic attraction between photogenerated electrons in the semiconductor-II CB and photogenerated holes in the VB of semiconductor-I, whereas in the Type II heterojunction, the electrostatic repulsion between the photogenerated electrons of semiconductor-II will prevent this from happening.

Z-Scheme System with Shuttle Redox Mediators

The redox mediated Z-scheme, also known as the PA-A/D-PS (first-generation) system, is made up of two semiconductors, PS I and II, and an A/D pair acting as an electron shuttle. Two set of charge carriers will delocalize in distinct semiconductors in such a system, causing water to divide into its component parts, H₂ and O₂. There is no physical contact between PS I and II, and the A/D pair controls the cascade type of vectorial electron transmission from PS II to PS I. The electrons from the CB of PS II will first reduce the electron acceptors into electron donors, leaving photogenerated holes accumulating in the VB of PS II, after the formation of electron–hole pairs as a result of photoexcitation in both PS I and II (Equation (23)). In the meantime, the photogenerated holes from the VB of PS I will transform the generated electron donors back into their oxidized state of being electron acceptors (Equation (24)). As a result, photogenerated electrons from PS II are indirectly shuttled to recombine with the help of an A/D pair, leading to effective isolation of the electrons and holes in PS I and II, respectively.

As a result, the system’s lowest VB and highest CB accommodate the photogenerated charge carriers, providing a sizable big overpotential for the Z-scheme reactions (Equations (25) and (26)) [165]. As such, it is possible to drive photocatalytic overall water splitting using this Z-scheme system while maintaining redox pairs’ generation.

A + e⁻ (CB of PS II) → D

(23)

D + h⁺ (VB of PS I) → A

(24)

2H₂O + 4h⁺ (VB of PS II) → O₂ + 4H⁺

(25)

2H⁺ + 2e⁻ (CB of PS I) → H₂

(26)

The red dotted lines in Figure 31 reveal that the electron A/D pair in the PS-A/D-PS system can also react with the photogenerated electrons and holes in the CB of PS I and the VB of PS II, which will inevitably result in a decrease in the effective number of charge carriers. As a result, the simultaneous evolution of H₂ and O₂ gas in the stoichiometric ratio attributed to the backward reaction is relatively challenging for a PS-A/D-PS Z-scheme system. With respect to this issue, several surface treatments are being implemented, such as loading metal co-catalysts, depositing rutile TiO₂, and exchanging Cs-H⁺; these are some of the attempts to hinder the backward reaction by impeding the adsorption of electron donors on PS II and electron acceptors on PS I.

At present, the redox mediators that are frequently utilized in the PS-A/D-PS Z-scheme system are Fe³⁺/Fe²⁺, IO³⁻/I⁻, [Co(bpy)₃] 3⁺/2⁺, [Co(phen)₃] 3⁺/2⁺, and VO²⁺/VO²⁺ [166,167]. Abe et al., for example, presented early work on the photocatalytic Z-scheme system. Anatase and rutile, two distinct forms of TiO₂, are interfaced via the IO³⁻/I⁻ redox pair as an ionic reaction mediator, serving as matching HEPs and OEPs [168]. Due to the wide range of valencies from −1 to +7, aqueous iodine is capable of acting as an electron mediator. Equations (27) and (28), however, demonstrate that I(0/1−) and I(1−/5+) are more stable during redox cycles in aqueous medium. The I₃⁻/I⁻ ionic pair is commonly used in dye-sensitized solar cells due to its facile reversible redox reaction and low photoabsorption, which only reaches about 500 nm [169]. However, because of the poor reactivity of I₃⁻ as an electron acceptor and the restricted selection of photosystems, the application of Z-schemes employing an I₃⁻/I⁻ redox cycle is highly limited [170]. The IO³⁻/I⁻ redox pair, on the other hand, has a standard potential that is close to the I₃⁻/I⁻ couple. Because of its electron nature, IO³⁻/I⁻ can thus be incorporated as an ionic redox mediator.

I₃⁻ + 2e⁻ → 3I⁻ E° versus NHE = +0.536 V

(27)

IO₃⁻ + 6e⁻ + 3H₂O → I⁻ + 6OH⁻ E° vs. NHE = +0.76 V

(28)

Figure 31. Schematic band energy diagram of the PS-A/D-PS Z-scheme system. Reprinted with permission from Ref. [169]. Copyright 2020, Wiley Online Library.

Figure 31. Schematic band energy diagram of the PS-A/D-PS Z-scheme system. Reprinted with permission from Ref. [169]. Copyright 2020, Wiley Online Library.

Even so, the IO³⁻/I⁻ redox cycle’s efficacy is greatly influenced by the pH of the medium. According to the Nernst equation [171], this ionic pair shows a higher contribution under increasing pH because of the shifting of the standard potential. Under UV light, Abe’s work involved the photocatalytic gas generation of a NaI solution containing Pt-loaded anatase TiO₂ (Pt-TiO₂-A1) as PS I and rutile TiO₂ (TiO₂-R2) as PS II [172]. At pH 3, the system exhibits a slow gas evolution that is attributed to the dissociation into I₃⁻ as the primary redox product, rather than IO₃⁻, which, as previously indicated, is not capable of acting as an effective electron acceptor in photocatalysis. However, the presence of favorable IO³⁻/I⁻ ionic couples, as opposed to I³⁻/I⁻, makes a basic solution of pH 9 the ideal environment for the IO³⁻/I⁻ redox system to work efficiently. The nonstoichiometric evolution of H₂ and O₂ is thought to be caused by the induction period of IO³⁻ production with a certain amount of I₃⁻, which is represented by the progressive increase in activity from pH 5 to 9. In addition, the development of inefficient I₃^— will also cause some shielding effect because of their large extent of light absorption. Therefore, it can be said that when the pH is higher than 9, the IO³⁻/I⁻ redox cycle can function effectively since there is no I₃ byproduct, leading to the stoichiometric evolution of H₂ and O₂ with high activity. Furthermore, in such a system, the ratio of IO³⁻/I⁻ has a major influence on the rate of gas evolution. The reverse reaction, also known as competitive oxidation, will result from the increased I⁻ concentration. For the PS-A/D-PS system to function with the IO³⁻/I⁻ redox mediator, an appropriate pH and NaI concentration are therefore essential.

Direct Z-Scheme Hydrogen Production

In order to build Z-scheme heterostructures using the boron-carbon nitride molecule, TiO₂ has been shown to be firmly connected. In comparison to pure TiO₂, the TiO₂/BCN nanostructure develops a higher concentration of photocatalytic H₂ generation activity and improved stability. The performance of the photocatalytic system is significantly influenced by the loading of BCN. For a 4% BCN/TiO₂ nanocomposite, the maximum photocatalytic activity was attained at a hydrogen evolution rate of 19.7 µmol h⁻¹g¹, which is two times greater as compared to using pristine TiO₂. The presence of middle sub-band single-electron oxygen vacancies with an innovative forbidden gap of TiO₂ has been found to be advantageous for the absorption of visible light in this sort of charge transfer mechanism. Due to the creation of an ohmic contact somewhere at the interface between BCN and new TiO₂, oxygen vacancies also benefit the Z-scheme charge transfer. Resultantly, the Z-scheme charge transfer in the BCN/TiO₂ nanocomposites can provide an outstanding TiO₂ oxidation ability and a strong BCN reduction ability [83].

10.4. Comparison of Type II Heterojunctions and Z-Scheme Heterojunctions

By using two semiconductors to produce Type II heterojunction nanocomposites, we can increase their charge efficiency and photocatalytic activity. The lifespan of photogenerated carriers can be increased by creating a heterojunction-based photocatalytic device that induces the separation of electrons and holes to two different sites [173]. The OER on O₂ evolution photocatalysts (OEPs) and the HER on H₂ evolution photocatalysts (HEPs) are two distinct photocatalytic reactions where gas evolution takes place. As shown in Figure 32, the migration of electrons and holes to a more electropositive CB and electronegative VB potential ascribed to the nature of charge transfer weakens the photocatalyic system’s redox ability even though heterojunction-type semiconductor nanocomposites are competent in facilitating charge separation.

In order to overcome these obstacles, the second method of photocatalysis, which is known as the Z-scheme photocatalytic system, introduces an anisotropic arrangement of two photocatalysts and an electron mediator to split water using a two-step photoexcitation process. The biomimetic artificial Z-scheme photocatalytic system, which draws inspiration from the natural photosynthesis of green plants, can concurrently meet the three demands listed below, which are lacking from both single-component photocatalysts and heterojunction-type nanocomposites: (i) an appropriate band-edge position with a sizable overpotential, (ii) small bandgap, and (iii) suppression of the recombination of electron–hole pairs [174]. Efficient water splitting can be achieved through Z-scheme photocatalysis, also referred to as a two-step photocatalytic system by the combined action of two isolated photocatalysts, one of which acts as the reduction site (photosystem I or PS I) and the other as the oxidation site (photosystem II or PS II). An electron transport chain known as a mediator is used to make it easier for electrons to flow between the two photosystems.

Z-scheme semiconductor composites, in contrast to heterojunction-type semiconductor composites, allow for a distinct electron flow profile that is ascribed to the inclusion of an electron-relaying channel. Electrons from the VB of PS I and II will be excited to the CB upon photoexcitation, leaving the photogenerated holes in the VB. The electron mediator will then move the photogenerated electrons in PS II to recombine the holes from the VB of PS I through ohmic contact [175]. Strong redox capabilities are maintained while electrons and holes can be accommodated into two distinct photocatalysts simultaneously, thanks to this unusual type of vectorial electron transfer. Thus, PS I is abundant with electrons for HER, whereas PS II is a hole-rich photocatalyst for OER. Resultantly, a Z-scheme system requires less change in Gibbs free energy to drive each photosystem than single-component photocatalysts and heterojunction-type composites [176]. It is evident from Figure 32b that the rational design of a Z-scheme provides effective charge separation in a separated position with a comparatively significant overpotential, which is more than enough to regulate the exceptional redox reaction.

10.5. Surface Plasmon Resonance (SPR) Effect

It was suggested that decorating the surface of TiO₂ with noble metal particles like Au, Ag, and Pt can enhance its photocatalytic activity [177]. Three factors contribute to the enhanced photocatalytic activity of metallic particle-coated TiO₂. First, by attracting the photogenerated electron and leaving a hole in TiO₂, metallic nanoparticles can boost the quantum yield. Second, metallic nanoparticles’ SPR effect can boost the ability of TiO₂ to absorb light and extend its sensitivity to visible light. The SPR effect is the cumulative rational oscillation of unpaired electrons on metal nanoparticles brought on by visible-light irradiation. Last but not least, SPR can strengthen the localized electric field around metal particles that interact with the coupled semiconductor and easily form electron–hole pairs in the semiconductor’s near-surface region [121].

It is well known that the LSPR (which is actually the collective vibration of the free electrons in metal nanoparticles) plays a key role in describing their optical properties. LSPR is a crucial component of biosensors because they are highly sensitive to the detection of molecularity-specific interaction analysis [178]. The aggregate vibrations with the interaction of resonant photons of metal nanoparticles of electrons like Au, Ag, and Pt are known as surface plasmon resonance. Often, plasmonic oscillation is limited to metal nanoparticles on surfaces with sizes between 10 and 100 nm [179].

LSPR, or the interface between such a conductor and a non-conductor, is where this phenomenon also occurs. Additionally, the LSPR behavior of a plasmonic material allows it to increase its absorption range into the visible region [180]. A near-field enhancement phenomenon called LSPR has the ability to capture light at the nanoscale with the highest optical absorption.

An appealing method for creating photocatalysts that can absorb visible light is to use the LSPR phenomena. Numerous studies have been conducted on the LSPR phenomena of noble metals like Au, Ag, and Pt [181]. Additionally, the compound photocatalyst’s capacity for light harvesting and the strength of the local electric field can both be enhanced with localized surface plasmon resonance of Au NPs. Plasmon-induced hot electrons can penetrate the Schottky barrier to enter semiconductors in Au–semiconductor nanocomposites, increasing the photocatalysis efficiency [182].

10.6. Co-Catalyst

One of the multiple functions of co-catalysts in photocatalytic reactions is to lower the rate of recombination [183]. This is accomplished in the case of metals by electron transfer to the co-catalyst, which happens because metals have a lower Fermi level than TiO₂ [184]. By encouraging the third stage of the water-splitting process, named as reducing the activation barrier of surface reduction and oxidation reactions, co-catalysts enhance H₂ evolution [185]. The mechanism of water splitting with both reduction and oxidation co-catalysts is shown in Figure 33. Pt was said to be a more successful co-catalyst for water splitting. Under UV irradiation, Maeda demonstrated stable overall water splitting into oxygen and hydrogen using rutile TiO₂ nanoparticles modified with Pt nanoparticles [186]. It is interesting to note that, as discovered by Sayama and Arakawa [187], Pt-loaded anatase TiO₂ photocatalyzed solely the H₂ production from water and not overall water splitting, under identical conditions. When co-catalysts are introduced to the surface of TiO₂, the bandgap can be reduced, allowing the photocatalytic activity to reach the visible light spectrum. The so-called surface plasmon resonance (SPR) effect is present in some metals. It is commonly known that Pt, Ag, and Cu exhibit a strong SPR effect [119].

Proton reduction into hydrogen and water oxidation to generate oxygen are the two half-reactions that normally make up water splitting into hydrogen and oxygen. The rate-limiting step between the two is water oxidation, since it requires four photogenerated holes. Furthermore, the water oxidation reaction is slower than the proton reduction reaction by approximately five orders of magnitude [188].

Ru is the least expensive element in the platinum group [189], so it is a desirable co-catalyst for industrial use. RuO₂ can reside on the base photocatalyst in both its oxide and metallic form. The metallic form of Ru enhances the hydrogen evolution reaction (HER) [190], while stimulates the oxygen evolution reaction (OER), thus contributing to the overall water-splitting reaction collectively [191]. On the other hand, compared to equally active Ir NPs, Ru nanoparticles have been demonstrated to be less stable in the long run [192]. Currently, Ir and its oxides are thought to be the most stable and effective catalysts [193]. As the rate-limiting step in the water-splitting reaction, IrO₂ functions as an OER catalyst [192].

While single metal co-catalysts might perform better in one of the two required reactions for water splitting, a combination of co-catalysts may work better for the overall effect [194]. Commonly, metal oxides such as RuO₂ and IrO₂ are considered as oxidation co-catalysts, while noble metals like Pt, Ag, and Pd are regarded as reduction co-catalysts [195]. One can further increase the hybrid material’s overall water-splitting activity by integrating both the reduction and oxidation co-catalyst on the same photocatalyst [196]. The literature has a number of studies on noble metal co-catalysts on TiO₂. The activity of Pt-Ir, Ir-Ru, Pt-Ru, and Pt-Ru-Ir combinations in nanoparticles deposited over rutile TiO₂ and anatase in an oxygen evolution reaction (OER) was examined by Fuentes et al. [197]. Tanaka et al. [198] synthesized TiO₂ by depositing large Au nanoparticles and smaller Pt nanoparticles and checked the H₂ production rate. Rutkowska and Kulesza [199] placed Pt-Ru nanoparticles onto TiO₂ supported on multiwalled carbon nanotubes for ethanol oxidation. Au, Pd, Pt, Ir, Rh, Ag, and Ru were deposited onto TiO₂ substrates by Lima et al. [200] in order to catalytically oxidize benzyl alcohol to benzaldehyde. Pt and Ru co-catalysts (alone and combined) were deposited by Caudillo-Flores et al. [201] onto Nb-doped TiO₂ for the photocatalytic production of H₂ from a methanol–water mixture.

Figure 33. Mechanism of water splitting with both reduction and oxidation co-catalysts. Reprinted with permission from Ref. [196]. Copyright 2019, Elsevier.

Figure 33. Mechanism of water splitting with both reduction and oxidation co-catalysts. Reprinted with permission from Ref. [196]. Copyright 2019, Elsevier.

11. Observation of Charge Carrier Kinetics in Heterojunction Structure

11.1. Transient Absorption Spectroscopy (TAS)

As mentioned in the introduction, non-radiative electron–hole recombination, which takes place before surface interactions between electrons and holes with water, is typically responsible for the moderate efficiencies of semiconductor water splitting. Transient absorption spectroscopy (TAS) is a widely used technique to assess the decay kinetics of carriers, since their lifetimes are on the scale of femtoseconds to nanoseconds, and even seconds after light absorption [202]. Although its application to PEC materials has been established, it is primarily utilized to study the kinetics of charge carriers in solar cell materials. A lock-in amplifier, a laser capable of producing femtosecond light pulses, and a device to analyze the absorption spectra as a function of time, wavelength, or applied bias make up the experiment’s setup [203]. Either the electron transient decay or the hole transient decay can be monitored with the appropriate scavenger. The technique by which an excitation (or pump) pulse promotes a fraction (0.1–10%) of the molecules to an electrically excited state is described by Barera et al. [203]. A weak probe pulse, which is one that is transmitted through the sample with a delay τ relative to the pump pulse, is one that is so weak that multiphoton/multistep processes are avoided during probing. The difference absorption spectrum can be calculated by subtracting the stimulated sample’s absorption spectrum from the ground state sample’s absorption spectrum (ΔA). A function of τ and wavelength λ, ΔA(λ,τ), is obtained by recording a ΔA spectrum at each time delay and changing the time delay (τ) between the pump and the probe. Consequently, since suppressing recombination and increasing carrier lifetimes are the primary goals of heterojunction systems, this is highly helpful for researchers studying solar fuel.

It was discovered, using nanocrystalline TiO₂ (nc-TiO₂) as an example, that the carrier lifetime was highly dependent on the pulse intensity and that recombination happens on the order of microseconds, while water oxidation takes place on the order of microseconds. Looking further, the absence of activity for water oxidation in N-doped TiO₂ under visible excitation was attributed to the photoholes created by the rapid decay of visible light, which happens on a much faster time scale than that needed for water oxidation as compared to nc-TiO₂ [204]. Pendlebury et al. used this method to study other visible-driven photoanodes and found that for α-Fe₂O₃, the amplitude of the long-lived hole signal is only ~10% of the initial hole signal. This suggests that most photogenerated holes still undergo rapid electron–hole recombination on a microsecond timescale, even under positive applied bias [205]. Subsequently, it was observed that the primary factor limiting water oxidation in α-Fe₂O₃ is recombination rather than surface kinetics [206]. As expected, the application of Co-Pi to the surface caused the observation of a cathodic shift in photocurrent as well as the development of long-lived hematite photoholes [207] as a result of the inhibition of electron–hole recombination.

The width of the space charge layer and Fermi level pinning decreased in the presence of surface catalysts, increasing the electron depletion layer’s size. When RuO_x particles were added to the surface of visible-light-activated Cu₂O photocathodes, the yield of long-lived (>100 μs) Cu₂O electrons increased significantly (as shown in TAS, Figure 34). This increase was attributed to a decrease in fast electron–hole recombination losses caused by hole transfer from Cu₂O to RuO_x, which increased the spatial separation of electrons and holes and aided in the photooxidation reaction by holes [208].

Ma et al. [209] recently reported on the correlation between the applied electrical bias and the yield of long-lived (0.1–1 s) photogenerated holes in BiVO₄. This correlation is attributed to the kinetic competition between water oxidation and the recombination of these surface-accumulated holes with bulk electrons across the space charge layer. Importantly, it was discovered that two mechanisms: firstly, fast (≤μs) electron–hole recombination, and secondly, the recombination of surface-accumulated holes with bulk BiVO₄ electrons, limit photocurrent generation in BiVO₄ photoanodes. For WO₃, Cowan et al. observed that quick (<μs) electron–hole recombination dominates in the absence of an electron scavenger, and water oxidation requires the production of long-lived holes with a lifetime in the milliseconds-to-seconds time scale [210]. Nonetheless, it is evident from the literature that this method should be applied to other junction architectures as it may be useful in demonstrating charge dynamics experimentally.

11.2. Electrochemical Impedance Spectroscopy (EIS)

Along with TAS, other methods that can investigate the kinetics of charge transfer and recombination are surface photovoltaic spectroscopy, photoluminescence spectroscopy (PL), and electrochemical impedance spectroscopy (EIS). Since radiative recombination, which is typically regarded as a minority phenomenon, is necessary for PL spectroscopy, it will not be covered in depth in this paper.

Applying a small amplitude sinusoidal AC voltage, V(t), and measuring the amplitude and phase angle (with respect to the applied voltage) of the resultant current, I(t), forms the basis of electrochemical impedance spectroscopy (EIS) [211]. The impedance (Z) can then be computed using Ohm’s Law, which states that Z = V(t)/I(t). The fraction of the hole current jh that is measured in the external circuit in EIS is determined by competition between the overall rates of hole transfer and recombination (Figure 35).

The rate constants k_t and k_r (s⁻¹) in the EIS response equations are used to represent the rates of recombination (cm⁻² s⁻¹) and hole transfer in terms of the surface concentration of “trapped holes”. It follows that k_r should be dependent on band bending, qΔΦ, since the electron concentration at the surface is given by

n_x=0 = n_bulk exp (−qΔΦ/k_B T)

where n_bulk is determined by the doping density.

By using this method, one can learn more about the phenomenological rate constants that describe the competition between recombination and charge transfer during the oxidation of water.

11.3. Surface Photovoltage Spectroscopy

This non-destructive method measures changes in band bending at the free semiconductor surface as a function of external illumination and can yield a wealth of qualitative and quantitative data for the heterojunction photocatalysis researcher. The surface photovoltage (SPV) is defined as the illumination-induced change in the surface potential. This includes, but is not limited to, the band position’s relative locations, defect states, surface dipole, diffusion lengths, recombination rates, the degree of band bending, and flat-band potentials.

12. Theoretical Modelling of Photocatalyst Junction Structures

There will be a discontinuity in the Fermi level at the interface when two materials with different Fermi levels first come into contact. In order to reach equilibrium, this will cause an electron transfer from the material with the higher Fermi level to the one with lower Fermi level. Moreover, the accumulation of a dipole layer at the interface between two materials due to the flow of electrons between them may further alter the respective band locations [213]. Therefore, in order to comprehend and forecast the behavior of semiconductor heterojunctions for photocatalysis, a number of physical mechanisms must be modeled. These mechanisms include the relative locations of the band edges of individual materials, which set the thermodynamic limits that establish the band offsets; the electronic structure of the materials, which provides details of the charge separation characteristics of photocatalysts such as electron and hole mobility; and the junction’s nature. This information can be used to calculate the band alignment, band magnitude, and impact of any interface dipoles present. Usually, these computations can only simulate systems up to a maximum size of a few thousand atoms [214].

Due to these size limitations, and since ab initio force calculations are usually most efficiently performed using a plane wave basis set, simulations typically concentrate on crystalline materials that are representative of bulk systems. Moreover, only ground state characteristics like cohesive energy and chemical bonding are precisely calculated using DFT. The accurate determination of excited state parameters is lacking, including the bandgap and optical absorption spectra. The lack of derivative discontinuity in the exchange correlation functionals prevents DFT from achieving the accurate determination of bandgaps, and bandgaps are reported in the literature as the simple difference in Kohn–Sham eigenvalues between the CBE and the VBE [215].

There have been a number of workarounds put in place to increase DFT bandgap accuracy. For metal oxide systems, two methods are frequently employed: hybrid functionals, which incorporate a fraction of the exact Hartree–Fock exchange to induce electron localization on particular sites, and DFT+U, which treats the localization problem of DFT195 by adding an empirical on-site Coulomb term +U [216]. The time-dependent Density Functional Theory method, also known as the GW approach, is often reserved for investigations involving clusters and will not be discussed in depth in this review.

Through photocatalytic water splitting, dye-sensitized photocatalytic (DSP) devices have shown promise as a sustainable method for producing hydrogen (H₂). Usually, a dye molecule is adsorbed on a semiconductor nanostructure and used as a photosensitizer in DSP systems. As illustrated in Figure 1, the dye molecules in the DSP system absorb incident light and introduce excited electrons into the semiconductor’s conduction band (CB). Reduction reactions, such as the reduction of water to produce H₂, can be accelerated by these excited electrons. With the help of the electron donor, the dye molecule can be renewed for cyclic usage [217].

Significant advancements have been made in the field over the past 20 years, and research has demonstrated that DSP systems can be created to cover the visible and even near-infrared portions of the solar spectrum by carefully choosing and matching the energy levels of semiconductors and dye molecules. Even with these tremendous advancements in the industry, DSP systems continue to have serious shortcomings, including low light-absorption efficiency [218]. It has recently been shown that dye molecules in dye-sensitized photocatalytic and photovoltaic systems can have their light-absorption efficiency increased by using the plasmonic Mie resonances of metal nanostructures, such as those made of gold (Au) and silver (Ag) [219]. Moreover, Mie resonance plays an important role in the selectivity and activity of the catalysts for energy efficient photocatalysis [220].

Under Mie resonance conditions, plasmonic metal nanostructures (PMNs) can show very high absorption cross-section values, up to five orders of magnitude higher than dye molecules. Thus, through a variety of mechanisms, such as the nanoantenna effect and plasmon-induced resonance energy transfer, PMNs (such as Ag and Au) can capture a significant portion of incident light, transfer the energy into the surrounding dye molecules, and increase their light-absorption efficiency [221]. In DSP systems, these plasmonic Mie resonance-mediated effects are used to speed up dye sensitization. But these PMNs-based DSP systems also come with built-in difficulties, like incompatibilities with traditional semiconductor fabrication, expensive material requirements, and commercialization problems that become more complicated because they require both metals and semiconductors.

Here, we suggest a different strategy, dielectric Mie resonance-enhanced dye sensitization, to address the previously noted problems. Figure 36 shows this strategy schematically. The Mie resonances can be found in dielectric (ε > 0) and plasmonic (ε < 0) materials with moderate (2.5–3.5) and high (>3.5) refractive index values, respectively [222].

Therefore, DSP systems based on metal oxide semiconductors with moderate-to-high refractive index values are most suited for the dielectric Mie resonance-enhanced dye sensitization shown in Figure 36. These metal oxide semiconductors include titanium dioxide (TiO₂), hematite iron oxide (α-Fe₂O₃), cerium (IV) oxide (CeO₂), and cuprous oxide (Cu₂O).

The dielectric Mie resonances of nanostructures of medium- and high-refractive-indexed metal oxide semiconductors can transfer the photonic energy into the adsorbed dye molecules via mechanisms like the resonance energy transfer, nanoantenna effect, and Mie resonance-mediated intense scattering effect, just like the plasmonic Mie resonances of PMNs. It is anticipated that the dielectric Mie resonance effects will improve the dye molecules’ light-absorption efficiency in the DSP systems (Figure 36c). One significant distinction is that a dielectric Mie resonance-based system (Figure 36c) does not require a separate light-enhancing material, in contrast to the plasmonic Mie resonance-based system (Figure 36b). This system’s primary benefit is that it only needs a dye and a dielectric semiconductor nanostructure.

In particular, the dielectric semiconductor nanostructure can function as both a source for obtaining excited electrons from the dye molecules and a Mie resonator for increasing the dye molecule’s light-absorption efficiency and promoting the reduction reaction (Figure 36c).

13. Conclusions and Future Perspectives

Hydrogen production is the best way to meet the present energy requirements, and there are various methods and sources to generate hydrogen. The best among them is solar water splitting because it is an economically viable and environmentally sustainable approach. Numerous semiconductors have been reported for solar water splitting, including titania, which shows many photocatalytic traits but faces a few limitations. A variety of modifications have been reported to upsurge its water-splitting activity, including doping, designing various heterojunctions, and sensitization. However, there are several challenges that must be addressed to enhance solar energy harvesting for improved kinetics of the water-splitting process. From our perspective, the following strategies should be implemented to improve the hydrogen production rate by TiO₂:

• Novel schemes should be developed that could maximize the light-harvesting capability of the TiO₂ which in turn could offer a higher water-splitting rate to yield hydrogen.
• Designing dual-catalytic TiO₂-based setups by mimicking the natural photocatalytic system is the best way to achieve sustainable energy. In this approach, another semiconductor is coupled with TiO₂ that modifies its electronic states and offers a suitable path for the enhanced kinetics of water splitting to generate hydrogen.
• The addition of a dopant tunes the bandgap of TiO₂, thus broadening its solar energy harvesting in the visible region, which is the first priority of the researchers for achieving a maximum energy input and minimal catalyst dose for the water-splitting reaction.
• Furthermore, a thorough understanding of the underlying reaction mechanism is a prerequisite to upsurge the catalytic proficiency of TiO₂ by improving its charge segregation, charge utilization, and reduced charge recombination rate.

Keeping in view the above modification strategies for titania, the more effective water-splitting setups along with highly efficient catalysts can be designed, which can prove their metal in augmenting the rate of solar-to-hydrogen conversion efficiency.