Chapter 1: Introduction to Photocatalytic Hydrogen Production

Sustainable, safe and clean energy supply is one of the technological problems addressed with greater intensity in recent decades by scientists, as energy consumption has grown significantly throughout the world.

To overcome the growing problem of pollution due to the use of fossil fuels, researchers have been attracted to the study of renewable energy resources, which can provide environmentally friendly alternatives that can help to minimize pollution problems.^1,2  Sunlight can be considered as the biggest energy source because its average daily energy (10²² J) and its utilization can not only meet the world’s annual energy needs, but also overcome the pollution problems that are arising due to the use of traditional methods of energy production based on fossil fuels.³ 

The conversion of sunlight into a usable form of energy (chemical, electrical or fuel production) has attracted much attention and is recognized as a form of green technology. In the past, for example, fuel cells have been developed to convert solar energy into electricity, and many improvements have been made to increase their efficiency, but large-scale energy production is still limited. Furthermore, regarding the transformation of sunlight into chemical fuels there are two possible methods: one is the reduction of carbon dioxide (CO₂) into hydrocarbons^4–7  and the other is the production of hydrogen (H₂).^8–10 

Hydrogen can be considered as a very efficient, clean and promising energy source.^11,12  In a fuel cell, only water (H₂O) is produced when H₂ reacts with oxygen (O₂). The importance of using H₂ as a fuel derives from the fact that it has an energy efficiency of 122 kJ mol⁻¹, which is higher than that of gasoline or any other fossil fuel.¹³  The industrial production of H₂ is still based on the use of fossil resources, such as oxidation of hydrocarbons, steam reforming of the gas and gasification of coal.^14,15  Although a significant amount of H₂ can be generated through these processes at a reasonable cost, the use of fossil fuels causes the release of greenhouse gases.¹⁶  Therefore, it is essential to establish a green, sustainable and effective approach for H₂ production.

Renewable energy (such as sunlight) can be used and H₂ production can therefore be recognized as a green alternative that can power everything from laptops to submarines.¹⁷  Compared to conventional methods based on non-renewable resources, recent research has revealed that the synthesis of H₂ using technologies involving sunlight is potentially competitive.¹⁸ 

An attractive method for hydrogen production is heterogeneous photocatalysis.¹⁹  A scheme of photocatalytic generation of H₂ in the water system is shown in Figure 1.1. This method is environmentally friendly and is gaining more and more importance in our globalized world.²⁰  In fact, it can be considered an independent procedure, which allows the conversion of solar energy into chemical energy. Photoreforming, on the other hand, which involves the presence of organic compounds as sacrificial agents, increases the yield of H₂ compared to the use of pure water. When the sacrificial compounds are oxygenated organic species that are bioavailable and sunlight is the source of irradiation, the process has virtually zero net greenhouse gas emissions. The source of the organic substrate has a large impact on the overall sustainability.²¹ 

Heterogenous photocatalysis has emerged as a viable approach to transform solar light into H₂ production, since 1972 when Fujishima and Honda²²  realized photoelectrochemical water splitting on a TiO₂ (rutile) anode, although the reaction proposed by the authors cannot be considered a photocatalytic reaction, strictly speaking. A photocatalytic process could be an economical way for the production of H₂ on an industrial scale, since often the synthesis of the photocatalyst and the design of the photoreactor present no particular difficulties.^23–25 

Obviously there are some drawbacks to consider regarding overall costs and efficiency, and only appropriate use of solar energy could make this technique cost-effective.

Very briefly, the three key steps that are typically involved in the photocatalytic production of H₂ are: (1) activation of the photocatalyst (generally a material with semiconductor properties), which consists in the formation of electron–hole pairs (e⁻/h⁺); (2) transfer of these charges to the surface of the photocatalyst particles; and (3) occurrence of oxidation and reduction reactions due to the presence of photoproduced charges (see Figure 1.2). It should be noted that not all couples give rise to reactive events, but some recombine, and their number per unit of time with respect to those that induce redox processes depends on the type of photocatalyst used. The recombination rate is an important parameter affecting the efficiency in all heterogeneous photocatalytic processes, especially those that are aimed at H₂ production.²⁶ 

“Band gap” is the term that describes the energy difference between the lower edge of the conduction band (CB) and the upper edge of the valence band (VB) of a semiconductor. Upon irradiation with light of energy higher than that of the band gap, the electrons in the VB move to the CB and the holes remain in the VB. Both ultraviolet (UV) and visible (Vis) light can be used depending on the semiconducting photocatalyst.

The e⁻ and h⁺ photoexcited in the presence of pure water give rise to H₂ and O₂, respectively, through the so-called water splitting mechanism. Gibbs free energy of 1.23 eV (237 kJ mol⁻¹) is required to convert H₂O to H₂ and O₂. Thus, for the photocatalytic water splitting reaction, the band gap energy (E_g) of the semiconductor must be >1.23 eV (or <1000 nm), and for activation to occur in the presence of visible light the energy must be less than 3.0 eV (or >400 nm).^27,28  Notably, the conduction band and the valence band edge energies must be compatible with the redox potential of the desired reactions.^19,29  The actual band structure of a semiconductor directly affects whether a particular reaction will occur (see Figure 1.3).²⁹ 

Both oxidation and reduction potentials of H₂O must be within the band gap of a semiconductor. The top edge of the VB must be more positive than the oxidation potential of O₂/H₂O (1.23 V vs. NHE, pH = 0), while the low edge of the CB should be more negative than the reduction potential of H⁺/H₂ (0 V vs. NHE, pH = 0).³⁰  In addition to the essential thermodynamic requirements concerning the potential values, many other parameters (many of which are related to the kinetics of the process) influence the photocatalytic production of H₂ by means of the water splitting reaction. We list here just a few examples, such as overpotentials, the charge transfer rate and the lifetime of photogenerated charges (obviously linked to their recombination rate as already mentioned above). The recombination of charge on the surface and in the bulk decreases the performance of the photocatalytic reaction by producing light and/or heat emission. On the other hand, a good separation of e⁻ and h⁺ (long lifetime) facilitates the production of H₂, as it favours the existence of more reactive sites.

Nevertheless, it should always be kept in mind that the values of potentials reported in the literature for the various materials are indicative because they often refer either to perfect crystals and/or to particular reaction environments (temperature, pH, type of medium and other variables). In reality, we work experimentally with photocatalysts that are not perfectly crystalline, often defective and impure of various species and, in the case of metal oxides, with different degrees of surface hydroxylation. Not to mention home-prepared (at different temperatures) composite materials, which are mixed or which support noble metals (for example, Pt) on the surface. For these materials, the mechanistic hypotheses that we researchers propose are often made taking into account purely theoretical thermodynamic values of band gap, VB and CB, which should instead be considered only indicative.

The photocatalytic production of H₂ is proportional to the amount of photogenerated electrons. To improve the efficiency of H₂ production, phenomena that consume photoexcited electrons must be avoided or reduced. Any process that increases the formation of electrons must be given great consideration; at the same time, to absorb more light a photocatalyst must have a small band gap and the reflection or scattering of incident light on its surface must be minimized. After formation of the photoexcited pairs, the efficiency of H₂ production in the water splitting reaction is strongly influenced by many factors, some of which have been cited in the previous section.³¹ 

Photocatalytic H₂ production from water in the presence of different organic compounds, i.e., sugar,¹¹  aromatic compounds,^32,33  biomass,³⁴  methanol,³⁵  ethanol,³⁶  ethylene,³⁷  amino acids,³⁸  fossil fuels³⁹  and lactic acids,⁴⁰  has been carried out in the presence of different semiconductors. It was noticed that a significant amount of H₂ was obtained through photoreforming reactions compared to that found with simple water splitting.

Most of the studies done using various organic compounds have been carried out by testing more or less sophisticated materials as photocatalysts, with the aim of improving their performance, but studies have also been concerned with mechanistic issues and the influence on the hydrogen production of the various organic compounds often used as sacrificial agents. Methanol is the most frequently used molecule as its simple structure allows an easier understanding of the reaction mechanisms. Of particular interest for the photoreforming process are some bioavailable substrates, such as glycerol, ethanol and sugars that come from biological substrates.⁴¹  Notably, these molecules, in addition to functioning as sacrificial agents, can form high added value compounds by partial oxidation, which occurs simultaneously and in tandem with the production of H₂. The type of sacrificial agent and the possible use of a cocatalyst are two particularly important factors for the photocatalytic performance. In the following section, some sacrificial agents are discussed in more detail.

For the production of H₂, methanol is considered an ideal molecule, and in early research it was assumed that it was the best and ideal feedstock for photoreforming. The photoreforming of methanol, in fact, produces formaldehyde and formic acid and, after further oxidation, only H₂ and CO₂. It is the most used sacrificial agent because it is the one that, with the same initial concentration, produces the greatest quantity of H₂, but no valuable compounds are obtained from its partial oxidation.^42,43  Jiang et al.,⁴²  during methanol photoreforming in the presence of commercial TiO₂ (P25) loaded with Pt, obtained an H₂ productivity of 3800 ± 130 mmol g_Pt⁻¹ h⁻¹.

Since the Gibbs free energy of the reaction has a negative value, the production of H₂ from saccharides, such as in glucose reforming, is thermodynamically advantageous. Hydrolysis of cellulose can provide glucose. Wastewater from food processing also has a high concentration of sugars. Consequently, a variety of catalysts for the photoproduction of H₂ using glucose as biomass have been investigated and the presence of Pt has been shown to be essential for good efficiency.^44–46  Bellardita et al. studied the photoreforming of glucose using different Pt–TiO₂-based photocatalysts, obtaining H₂ and high added value products at the same time.¹¹  The anatase, brookite and rutile TiO₂ samples showed different photoactivities giving rise to a different distribution of the partial oxidation products. Rutile was more efficient than anatase in converting glucose, by producing arabinose and erythrose as partial oxidation products. Pure brookite showed the same behaviour as rutile. Gluconic acid formed mainly in the presence of anatase. The different reactivity has been tentatively attributed to the formation of ˙OH radicals on the surface of anatase and of peroxidic species on rutile. Rutile and brookite were the most active photocatalysts for H₂ formation, with approximately 1700 µmol of H₂ produced in 7 hours of irradiation in the presence of Pt–brookite. The presence of Pt proved to be essential for the production of H₂.

It is very interesting to use glycerol as a sacrificial agent to produce H₂ through the photoreforming reaction. Constituting approximately 10% of by-products from the biodiesel manufacturing industry, glycerol is produced and accumulated in large quantities. Consequently, it is considered a waste and its valorisation by producing H₂ and compounds with high added value with the photocatalytic method under green conditions is desirable. Several photocatalysts have been tested together with glycerol, but TiO₂ is the most popular one.^47–50  Using Cu₂O–TiO₂ as photocatalyst, Pecoraro et al. obtained H₂ and CO₂ as the main gaseous products together with 1,3-dihydroxyacetone and glyceraldehyde, from photoreforming in the liquid phase.⁵¹ 

Ethanol, which can be produced on an industrial scale from renewable biomass (cellulose or lignocellulose), has been widely proposed as a sacrificial agent for H₂ production.

From the photoreforming of ethanol, the products formed are acetaldehyde and acetic acid, which, however, can undergo further oxidation depending on the reaction times and the experimental conditions used. Some other products, i.e., carbon dioxide (CO₂), ethane (C₂H₄), methane (CH₄), carbon monoxide (CO) and ethylene (C₂H₆), have also been found.^52–54 

It is difficult for a single-phase photocatalyst to exhibit both good redox capability and a broad light absorption spectrum at the same time. Furthermore, in single-component photocatalysts, there is the possibility of recombination of the photogenerated charges, with a consequent reduction in the efficiency of the process.⁵⁵  Therefore, the performance of single-phase materials has been greatly improved using techniques including doping, morphological or crystal lattice control, surface modification, sensitizing and coupling of semiconductors with appropriate band energy position.^56–60  Based on the band structures of photocatalyst heterostructures, there are four main types of heterojunctions (type I, type II, Z-scheme and S-scheme).

In type I heterojunctions, the VB of semiconductor 2 is above the VB of semiconductor 1, while the CB of semiconductor 2 is lower than the CB of semiconductor 1 (see Figure 1.4). After irradiation of the heterostructure, the electrons move from the CB of semiconductor 1 to the CB of semiconductor 2, while holes move from the VB of semiconductor 1 to the VB of semiconductor 2. Therefore, semiconductor 2 becomes enriched in electrons and holes.⁶¹  However, the charge carrier separation only improves slightly since electrons from the CB of semiconductor 2 and holes from the VB of semiconductor 2 can still recombine easily. It is important to remember that the prerequisite for charge relocation is the creation of an internal electric field that favours not only their separation, but also their transfer to reacting adsorbed species.^62,63 

In type II heterojunctions, both the positions of the CB and VB of semiconductor 2 are lower than those of semiconductor 1 (Figure 1.5). Electrons move from the CB of semiconductor 1 to that of semiconductor 2, while photogenerated holes move from the VB of semiconductor 2 to the VB of semiconductor 1. Also, the bending of the band at the junction interface is affected by the chemical potential difference between the two semiconductors. The electric field established by band bending favours the displacement of photoexcited electrons and holes to opposite sides, resulting in effective charge separation and greatly increased photocatalytic activity. Holes accumulate in the more positive VB and electrons in the less negative CB of the photocatalyst, weakening the reduction and oxidation capabilities.^62,63  However, different type II heterojunction systems have been used for H₂ formation.^64–67 

A Z-scheme heterojunction system is shown in Figure 1.6. After irradiation, the electrons of the conduction band of semiconductor 2 recombine with the holes of the valence band of semiconductor 1. Thus, electrons with high reduction capacity are available in the CB of semiconductor 1 and holes with high oxidation power in the VB of semiconductor 2. The artificial Z-scheme heterojunction structure has received much attention, and a number of Z-scheme structures of various types have been created to make some photocatalytic reactions more efficient. The disadvantage of type I and type II heterojunctions is that at each active site the light-induced electrons and holes lose some of their redox capacity after migration. The Z-scheme heterojunction system has the advantage not only of ensuring an efficient spatial separation of the electron and hole pairs, but also of producing electrons and holes with remarkable reducing and oxidizing power, respectively.

Nevertheless, the Z-scheme heterojunction system also has some problems to face. One problem is that the photoreduction and photooxidation reactions can influence the behaviour of the photocatalysts for which the more or less marked accumulation of electrons and holes takes place. The other problem is the occurrence of light-induced recombination of electrons and holes, which, although efficiently separated by light, causes them to be consumed in significant quantities. However, the Z-pattern heterojunction system is attractive due to the particular type of electron transfer.^62,63  The direct Z-scheme system is useful towards H₂ production when the potential of the CB edge of semiconductor 1 is more negative than the H⁺/H₂ redox potential.^68–70 

The S-scheme heterojunction system (Step-scheme) is illustrated in Figure 1.7. Considering the level of the energy bands, the two coupled photocatalysts involved can be classified as an oxidation photocatalyst (OP) and as a reduction photocatalyst (RP).^71–73  The first photocatalyst presents lower CB and VB positions, and consequently a lower Fermi level, than the second one. The work function of the OP is larger than that of the RP.

When the OP and RP are in close contact, transfer of the photoproduced charges occurs according to the following steps.

1. After simple contact between the two materials, the electrons will move spontaneously from the RP with a lower work function to the OP, until their Fermi energy levels (E_f) reach equilibrium (Figure 1.7(b)). This results in different charges on the two sides of the contact surface, with negative and positive charges on the OP and RP sides, respectively. Thus, an internal electric field (IEF) is formed.

2. Under irradiation (Figure 1.7(c)), the IEF formed accelerates the electron transfer from the OP to the RP. Furthermore, its formation enhances the bending of the energy bands: downwards in the OP and upwards in the RP. At the interface, the photogenerated e⁻ in the CB of the OP and h⁺ in the VB of the RP recombine under the coulombic attraction. Thus, electrons and holes in the OP and RP, respectively, are consumed, but electrons with strong reducing ability in the RP and holes with strong oxidizing ability in the OP remain. Consequently, the S-scheme heterojunction photocatalyst with such a mechanism has strong oxidation and reduction abilities^62,74  and thus can be used for effective H₂ production.

The photocatalytic activity towards H₂ production can be evaluated directly from the amount of H₂ obtained in the specific time interval under the irradiation source. Different photocatalytic setup configurations and irradiation sources used by different research groups show different photocatalytic activities even for the same photocatalyst. As has been reported in the sections above and as will also be explained later in this chapter, there are many methods that are proposed to improve some surface, structural and morphological characteristics of the photocatalysts or to modify the experimental conditions, in order to obtain more significant production of H₂. However, the results are often presented in an uneven way and it is not easy to make comparisons. For example, some authors refer to the surface area of the photocatalyst, others to the weight and the duration of the tests varies from one paper to another. Unfortunately, the quantities of H₂ produced photocatalytically still do not seem to be attractive from an industrial point of view and it is worthwhile to conduct further research.

It is important that a good photocatalyst has high stability to produce H₂. To verify the stability it is necessary to carry out long experiments and to repeat them several times using the same photocatalyst recovered from the reaction environment. Furthermore, after the catalytic tests, the catalyst should be characterized in detail to verify any changes due to (photo)corrosion phenomena.

Steady increase in product concentration or steady-state reaction rate during prolonged irradiation in batch or turbulent-flow reactors are indicators of catalyst stability. Photocorrosion is the main cause of the poor stability of many metal sulfides when used as photocatalysts. CdS, for example, is subject to anodic photocorrosion during the formation of H₂.⁷⁵ 

Many types of semiconductors have been designed and used in the photoassisted formation of H₂, but these have not yielded significant efficiency to justify scale-up of the process, despite some encouraging results. Due to its non-toxicity, low cost and great chemical (photo)stability, TiO₂ has been widely studied and used (see Section 1.8). However, this material has a fairly high band gap (3.2 eV for anatase), absorbs light only in the near-UV range and exhibits a relatively high recombination rate of the photogenerated charges.⁷⁶ 

After the initial studies on obtaining H₂ and O₂ by water photosplitting using TiO₂ under UV irradiation,⁷⁷  other research followed with various semiconductors, such as CdS and WO₃, with band gaps that allow the use of visible light. As mentioned in the previous section, the drawbacks that these types of materials present have prompted researchers to find other solutions to improve the efficiency of photocatalytic water splitting. To obtain active photocatalysts under visible light irradiation, some approaches have been proposed, such as (i) doping of metallic and non-metallic ions; (ii) emerging solid solutions; (iii) sensitization by dyes; and (iv) development of new single-phase photocatalysts.²⁷ 

Metal ion doping is known to be the most efficient method of producing impurity levels in the band gap necessary for the creation of visible light-active photocatalysts. This method, which has been known for many years, activates wide band gap photocatalysts in the visible light region. Much work has been devoted to the development of metal ion-doped broad-band photocatalysts. Some examples of semiconductors that have been doped are ZnS,⁷⁸  SrTiO₃,^79,80  TiO₂^81–83  and La₂Ti₂O₇.⁸⁴ 

Another frequently used method that modifies photocatalysts by reducing their band gap and increasing photocatalytic activity under visible light irradiation is doping with non-metal ions. Non-metal ion dopants, unlike metal ion dopants, tend to develop donor levels in the forbidden band gap, as well as shifting the valence band edge upwards. Some UV light-active oxide photocatalysts, such as Ti-based oxides,^85–87  Ta,^88,89  Zr,^90,91  Nb^92,93  and Ta-based oxides, have been modified using non-metal ion doping technology.

In recent years semiconductors co-doped with metal–non-metal ions have shown high photocatalytic performance when irradiated with visible light. TiO₂ has often been modified by co-doping with metallic and non-metallic ions to enhance its photocatalytic performance, especially for the removal of pollutants. Furthermore, it was found that modified SrTiO₃, co-doped with La and N species, showed higher photocatalytic performance under visible light than unmodified SrTiO₃.^94,95  Very few attempts have been made on the use of photocatalysts co-doped with metallic and non-metallic ions and irradiated with visible light for the formation of H₂ from H₂O. Nitrogen-doped TiO₂ powders doped with various metals, such as Fe, Cr, Ni and Pt, have been used for photocatalytic water splitting under visible light. Among these the best was Ni–N–TiO₂, which showed good stability and produced 490 µmol g⁻¹ h⁻¹ of H₂ in the first 6 hours of irradiation.⁹⁶ 

Another interesting technique to control the band structure of the semiconductor photocatalyst is the development of solid solutions between broad-band and narrow-band semiconductors. In a solid solution, by modifying the proportion of wide- and narrow-band semiconductor components, the extent of the band gap and the position of the bands can both change. ZnS and CdS, broad-band and narrow-band semiconductors, respectively, can be combined to create solid solutions of Cd_xZn_1−xS sulfide, which have attracted the attention of researchers for applications using photocatalysis in the presence of visible light.^97,98  Cd–Zn–S solid solutions containing a high amount of surface defects display an H₂ evolution rate of 5.2 mmol h⁻¹ g⁻¹ under visible light irradiation without loading of noble metal.⁹⁸ 

Another effective method of making wide-band semiconductors active under visible light is by dye sensitization.^99–101  Excitation of the dye and the subsequent movement of charge from the dye to the semiconductor is the basic principle of dye-sensitized solar cells used in water splitting reactions. Since the electron transfer process is similar to that of dye-sensitized solar cells, dye-sensitized semiconductors can be used for water splitting when exposed to visible light.¹⁰² 

One renewable, and essentially unlimited, source of energy is sunlight. Consequently, the use, transformation and/or storage of solar energy is a popular goal in place of competing renewable or non-renewable sources. In this regard, scientists around the world consider photocatalysis to be a fascinating scientific method for producing fuels by absorbing solar energy. It has been demonstrated that the photoreforming of oxygenated compounds derived from biomass (see Section 1.3.1) in the presence of some photocatalysts can occur easily and with good efficiency under visible-light irradiation.^103–106 

The Arrhenius equation describes how the rate of photoreforming changes with temperature according to the activation energy of the limiting kinetic step. The movement of charge from the active sites on the surface of the irradiated material and the adsorbed species in the elementary steps of a photocatalytic reaction is a process driven by the absorption of light.^107,108  The dual use of solar radiation through both the effective use of light by the semiconductor and its local heating (photothermal process)^109,110  is a process that offers evident advantages for the production of H₂, especially taking into account improvement in the efficiency of the photoreforming by increasing the temperature. It should also be noted that the performance of the photocatalytic processes under examination is considerably improved using reactors that are suitably designed and with particular configurations and symmetries with respect to illumination (see Section 1.7.10).¹¹¹ 

The photocatalytic activity can be influenced in various ways by the acidity or basicity of the solution used. In fact, the processes are affected by the pH value of the medium, which can also significantly influence the energy position of the semiconductor bands, the redox potentials of the molecules, the speciation of the substrate and the extent of its protonation (for acidic or basic species), the surface charge of the semiconductor used, the electrokinetic potential (zeta potential, ζ), the surface adsorption of the species involved in the reaction, the stabilization of reaction products or intermediates and the aggregation of photocatalyst particles. While the surface reaction and charge transfer depend on the type of interface and, therefore, also strongly on the pH, the bulk structure of a photocatalyst is unlikely to change significantly as a function of pH.⁴¹ 

Substrate concentration has a large impact on H₂ production. As substrate concentration increases, the rate of H₂ production increases to a certain point, but then begins to decrease as excessive amounts of substrate can compete with H₂O adsorption.⁴⁶ 

Water is typically required for the stoichiometric synthesis of H₂ in the photoreforming of oxygenated compounds. Consequently, the water/substrate ratio is crucial for the selectivity of the process, although it is possible to carry out some reforming processes in the absence of water, such as the dehydrogenation or the decomposition of formic acid. It should also be noted that the water on the photocatalyst surface can act as an oxidant, forming intermediate hydroxyl radicals, or function as a proton carrier, thus influencing the rate at which the reaction occurs.^112,113 

The amount of the photocatalyst, as in any heterogeneous photocatalytic reaction, is important and influences many parameters relating to photoactivity, such as the reaction kinetics and the yield and selectivity towards the products. It is necessary to establish experimentally the optimal amount of photocatalyst for the photoreactor used, taking into account not only the volume of liquid used, but also the geometry of the photoreactor. Light scattering, transmission and reflection should be minimized, while ensuring that all or most of the incident photons are absorbed by the reacting system. In cylindrical photoreactors, for example, a commonly used methodology is to gradually add the powdered photocatalyst to the aqueous reacting system and measure the transmitted light with a radiometer until it reaches very low values close to zero. In this way it can be assumed that most of the photons remain in the reacting system and give rise to formation of e⁻/h⁺ pairs, which induce the reaction or recombine, generating heat. Obviously, determining the optimal amount of photocatalyst is not easy for other photoreactor geometries.^114–117 

Many catalysts are inactive to produce H₂ in the absence of a cocatalyst. Noble metal nanoparticles have been considered excellent cocatalysts for H₂ photoproduction. A widely used way to improve photocatalytic efficiency is to load the cocatalyst on the surface of a photocatalyst, promoting charge separation and providing more reaction sites. In fact, due to the difference in Fermi levels between the semiconductor and the noble metal nanoparticles, a Schottky barrier is formed and the electrons move from the semiconductor CB towards the noble metals.¹¹⁸ 

Using an optimal amount of cocatalyst is important for the photocatalytic production of H₂ and depends on the cocatalyst used, as well as the surface area available on the photocatalyst particle. By increasing the amount of cocatalyst present on the photocatalyst surface, the photocatalytic efficiency increases but at some point it then starts to decrease. This is due to the increase in the dimensions of the particles, the unsatisfactory dispersion on the surface (accumulation phenomena can occur with an increase in the dimensions of the cocatalyst nanoparticles) and the shielding effect of the active sites of the photocatalyst.

The noble metals (Pd, Ru, Pt) are considered the best candidates for their broad work function.⁹  Cu₃P cocatalyst on the surface of g-C₃N₄,^119,120  metal borides and carbides, due to their semi-metallic characteristics,^121,122  were also used to produce H₂. Notably, the cocatalysts NiS and CuS played a crucial role in the production of H₂ on g-C₃N₄.^8,123,124  To improve the efficiency in the production of H₂, co-loading of reduction and oxidation cocatalysts has also been extensively studied, which has improved effects in the separation and movement of the charges.¹²⁵  Some cocatalysts are represented in Figure 1.8.

Semiconductor coupling is another useful method for enhancing photoactivity through efficient photoinduced charge separation. As mentioned above with regards to CdS (Section 1.5.2), the photoinduced holes in the VB cause a significant auto-oxidation making the material extremely unstable towards photocorrosion. To improve the photocatalytic performance and stability CdS has been coupled with other photocatalysts with various band gap values, such as TiO₂, ZnO, KNbO₄ and LaMnO₃. In the CdS–TiO₂ coupled material, with irradiation under visible light the photoinduced electrons in the CdS migrate towards TiO₂, while the holes remain in the CdS. This obviously increases the photoactivity, facilitating charge separation, preventing recombination and making more electrons and holes available for the adsorbed reacting species.^126,127 

The crystal structure is an important factor influencing the photoactivity. In the case of TiO₂, anatase is generally more active than brookite. In anatase TiO₂, the oxygen vacancies can trap the photoexcited electrons and it is easier for them to reach the Pt particles present on the surface, which act as a trap and which are used in the case of H₂ production. The low activity of rutile TiO₂ can be attributed to the fact that the photoproduced electrons are trapped in intrinsic structural defects.^128–130  Surprisingly, brookite turned out to be very active in the production of hydrogen.^10,131  Mixed phases of TiO₂ were found to be more active than pure phases of TiO₂.^132–134  To produce H₂, the transfer of electrons at the phase junction is important. While lattice defects act as recombination centers for photoexcited electron–hole pairs with decreased photocatalytic performance, good crystallinity significantly improves electron transport.¹³⁵ 

In general, particle size is a fundamental aspect in electron–hole recombination and charge transfer, which is mainly controlled by the quantum confinement.⁸¹  Nanomaterials have been widely used in heterogeneous photocatalysis. Decreasing particle size is beneficial for photocatalytic performance as electron–hole recombination is reduced relative to the bulk material.^136–138  Reducing particle size increases surface area and provides more surface sites. High photocatalytic performance has been observed in the water splitting reaction with smaller particle size and higher surface area for NaTaO₃.¹³⁹  The reason for this is the increased surface interaction capacity of electrons and holes instead of mass recombination.¹⁴⁰ 

In almost all the articles described in this chapter the photoreforming reactions of organic species for the production of H₂ have been carried out in a batch photoreactor.¹⁴¹  In general, the photoreactor used should have some characteristics that are necessary to correctly obtain the concentration data of both the substrate in the liquid phase and the H₂ in the gaseous phase.¹⁴²  First, the reactor volume should be sealed during the reaction to avoid ingress of air, which could decrease the efficiency of the reduction reaction. Furthermore, the exchange of gaseous species between the reactor and the external environment could compromise the correct determination of the H₂ developed. For this reason, all the ports in the reactor should be equipped with valves built with inert materials. Furthermore, a common feature of all photocatalytic devices is the correct positioning of the radiation source with respect to the reactant volume. This is very important, to have a quite uniform distribution of the radiant field within the reactant volume. For irradiation systems placed outside the reactor there are no particular problems, while in the case of lamps immersed directly in the reactant volume it is necessary to seal the casing containing the reactor and the lamp.

Another important characteristic of the reacting system is the presence of a head space, which allows accumulation of the gaseous species produced during the reaction. At least three ports are required at the top of the reactor: the first to allow an inert gas to bubble into the liquid dispersion to purge O₂ present in the head space and dissolve in the reacting solution; the second and third from which gas and liquid can be withdrawn, respectively (see Figure 1.9). As far as continuous reactors are concerned, what has been said previously regarding tightness remains valid. However, continuous systems often do not allow accurate chromatographic analysis of H₂ due to its low productivity. Therefore, it is necessary to provide an accumulation volume of H₂ at the outlet, which allows it to be separated from the liquid phase.

TiO₂ is the most widely used material as a heterogeneous photocatalyst and constituent of photoanodes (see Figure 1.10).¹⁴³  There are many in-depth reviews on TiO₂ used for the production of H₂.^27,144  The most common form of titanium dioxide is anatase, which behaves as an n-type semiconductor.¹⁴⁵  Crystalline TiO₂ nanoparticles with dimensions <10 nm are more efficient in heterogeneous photocatalysis than amorphous TiO₂. The latter, in fact, shows a higher photogenerated electron–hole recombination rate and obviously does not benefit from the quantum size effect. TiO₂ still has limitations for its large-scale application in photocatalysis due to the large band gap value, which does not allow its activation under visible light irradiation, which constitutes about 50% of the solar spectrum, and its relatively high recombination rate of the photoproduced pairs.¹⁴⁶  Only 4–5% of UV photons present in sunlight can activate TiO₂. However, TiO₂ nanoparticles continue to be the most-studied nanomaterial for H₂ formation.¹⁴⁶ 

Some of the semiconductors used for the photocatalytic production of H₂ show different photocatalytic performances (see Table 1.1). For example, it has recently been demonstrated that for a number of commercial samples the activity is due to various factors, such as surface area, morphology, crystallinity, phase type and particle size. In order to increase the photocatalytic efficiency it is desirable, in principle, to have photocatalysts with large surface area and high crystallinity; in fact, in this way the recombination of electrons and holes is reduced and generally increases the number of surface sites, even if the increase of the surface area at the expense of the crystallinity does not always guarantee this fact.¹⁴⁷ 

The three main polymorphic forms of TiO₂ are anatase, brookite and rutile. Numerous synthetic techniques are known in the literature for the production of TiO₂, which can also be obtained in the form of various types of polymorphs as a mechanical mixture or mixed material (when there is an interface between particles of one polymorph and another).¹⁴⁸  Anatase, which has a tetragonal crystal configuration with each octahedron sharing corners to generate planes (0 0 1), is the most widely employed polymorph in heterogeneous photocatalysis. Rutile, which on the other hand is mainly used as a white pigment, even if its use in photocatalysis and photoelectrocatalysis is quite frequent, has a tetragonal crystalline configuration made up of octahedra sharing the edges that form the planes (0 0 1). Angular and edge-sharing octahedra form the orthorhombic crystal structure of brookite. Anatase is the polymorph that generally gives the best performance for the photocatalytic production of H₂ because it has better stability, the photoproduced pairs are better separated and its band gap is slightly smaller.^149–151 

However, it must be taken into account that in some cases photocatalysts made up of mixed phases can also be used, and the level of photoreactivity depends on the experimental conditions used. Ultimately, it is not easy to establish which is the best-performing TiO₂ phase or phase mixture.

In this chapter the main principles on which heterogeneous photocatalysis is based have been reported, considering its possible use for the formation of H₂. The information provided does not purport to be exhaustive, but only a useful guide for the reader, who will find in the references chosen, among the many in the literature, a way to delve deeper into individual aspects of greatest interest to them.

It can be concluded that, despite research efforts, photocatalysis has not been able to produce H₂ at an industrial level, although its use is desirable from an environmental point of view. However, it seems useful to continue studies on the formation of H₂, especially in the presence of sacrificial waste substances and in particular with biomass that can be transformed into high-value molecules.

Furthermore, it has been shown that many parameters can be modified in an attempt to improve the photocatalytic efficiency and therefore the quantities of H₂ formed, which to date have been rather low.

It is mandatory to use safe materials, such as the photocatalysts, and to avoid those that exhibit leaching or photocorrosion whether in powder or in film form. The presence of heavy metals as dopants should be avoided when aiming for a practical application, although in some cases a significant improvement in efficiency has been noted with their use. The most used type of photocatalyst for the production of H₂ has been TiO₂ in its various polymorphic forms (also coupled with other materials) containing Pt or other noble metals on the surface. Although the cost of Pt is high, the small quantity needed could justify the use of Pt-containing photocatalysts in the presence of high production. Another strategy, however, could be to replace the noble metal with inexpensive materials. In conclusion, despite the difficulties to overcome, the photocatalytic production of H₂ appears to be a challenge and a fascinating prospect. It is well known, in fact, that its use as a fuel could contribute significantly to reducing today’s air pollution.