A review of renewable energy generation using modified titania for photocatalytic water splitting

As the human development index relies on per capita energy consumption,¹ it is important to use the abundantly available clean solar energy for securing future energy requirements. By adopting this approach, we can reduce our dependency on fossil fuels, which are the main sources of greenhouse gases.² Hydrogen is an ideal clean fuel that can be obtained from water, which is an abundantly available renewable resource. Hydrogen energy economy has already been set in progress, but it is much expensive than fossil fuels due to the costly production methods such as steam and plasma reforming oxidation and electrolysis.³ Presently, in commercial production, 5% renewable energy contribution comes from hydrogen generated through water electrolysis.⁴ Since the first discovery of TiO₂ based photoelectrolysis of water in 1972,⁵ researchers have focused on hydrogen generation via photocatalytic water splitting using a variety of semiconductors such as TiO₂,^6,7 graphitic-carbon nitride,^8,9 and CdS.^10,11 Among these, TiO₂ is considered as a benchmark photocatalyst owing to its chemical stability, suitable bandgap, and comparatively lower cost.^12–14 However, it suffers from two limitations; inefficient absorption of visible light (because of 3.2 eV bandgap) and the relatively high rate of recombination of electrons and holes leading to lower yield of hydrogen production, i.e., 0.7% during 5.5 h.¹⁵ Several strategies have been conceived and tested for improving the visible light activity of titania, but until now no available photocatalyst has so far satisfied the requirements of absorbing visible light radiations, stability, cost affordability, and efficient hydrogen generation.¹⁶ Thus, to address these difficulties, it is extremely desirable to search for new materials that could split water by absorbing visible light. In this regard, modifications in titania for accelerating photocatalytic hydrogen generation from water is a favorable approach. Though extensive literature is available on the photocatalytic water splitting performance of titania, no review article is available on modified titania that could offer details of various influential factors leading to enhanced photocatalytic water splitting. Herein, we present the influence of energy bandgap, surface morphology, and sacrificial reagents that affect the splitting of water along with the recommendation of enhancing the performance of titania by the addition of electron donors, doping with metal and nonmetal ions, and coupling with composite semiconductors.

Water splitting is a multi-electron process^17,18 as shown in the following equation:

The energy required to generate one molecule of H₂ from water is 2.458 eV, involving a potential of 1.229 V.¹⁹ A growing body of literature has summarized seven stages of H₂ production in four essential steps as shown in Fig. 1. The process involves light irradiation (stage 1), electron absorption (stage 2), (e⁻/h⁺) migration (stage 3, 4, and 5), and reduction and oxidation half reactions (stage 6 and 7).²⁰ 

In the first stage, light absorption is affected by the morphology of the photocatalyst and can be enhanced through hierarchical order macroporous or mesoporous structures (due to the scattering effect). Working in this area, Wang et al. synthesized hierarchical meso-/macroporous TiO₂ nanostructures (HOMMTs) and found that the resultant HOMMTs display good photocatalytic efficiency possibly due to ordered mesopores (11 nm) connected to macrospores (470 nm–1350 nm) and a large surface area.⁷ The electronic excitation of the semiconductors (stage 2) relates to their electronic structures. In this connection, Zhu and Zäch reported that the bandgap separates the valance bands (VB) of semiconductors from their conducting bands (CB).²¹ For water splitting, the semiconductor bandgap should possess a value greater than 1.23 eV. Therefore, transferring electrons to CB and generating positive holes in the valence band by irradiating TiO₂ with energy greater than bandgap energy (E_g) leads to the formation of (e⁻/h⁺) pairs [Eq. (2)]. Due to charge separation (stage 3), two unfavorable stages take place related to (e⁻/h⁺) recombination in the bulk (stage 4) and at the surface (stage 5). Li et al. reported that the charge recombination can be reduced by applying electric fields.²² Finally, both photoexcited electrons and holes move freely and can absorb on the semiconductor. The shorter relaxation time of the CB as compared to the time required for crossing the bandgap results in an internal energy level equilibrium.²³ Only the e⁻ and h⁺ transferred to the semiconductor surface are accountable for surface reaction and can be trapped by co-catalysts to stimulate photoreduction [Eq. (3)] and photo-oxidation reactions [Eq. (4)], represented by stage (6) and stage (7), respectively,

The performance of semiconductors depends upon bandgap and the nature of radiations passed. The bandgaps of typical photocatalysts can be seen in Table I.

The wavelength of the absorbed light decreases with bandgap widening; hence, a semiconductor is considered as UV active if its bandgap exceeds 3.15 eV, and active to visible light if its bandgap remains lower than 3.15 eV. To achieve the maximum photocatalytic activity, the bandgap width must be in the range of 1.23 eV–3.26 eV,³¹ and the potential for reduction and oxidation reactions should be within the bandgap limits of the photocatalysts. Working in this area, Phoon et al. reported that the value of the conduction band should be more negative as compared to the cathodic potential (−0.41 eV at the NHE scale in a medium of pH 7) of hydrogen, while the valence band should have a more positive value than the oxidation potential (+0.82 eV at NHE scale in a medium of pH 7).³² A semiconductor of lower CB value reduces its reduction capability and enhances its combination with other more negative CB semiconductors, allows electrons to transfer easily to the adjacent semiconductor and, thus, reduces H⁺ to H₂ without (e⁻/h⁺) pair recombination.³³ 

The main objective of modifying photocatalysts is to improve their visible light activity toward water splitting. Photocatalyst modification can be carried out by a number of ways as discussed in Subsections III A–III C.

Guo and Ma claimed that adding electron donors (sacrificial reagents) to titania can enhance photocatalysis as the photogenerated holes are replaced by the sacrificial reagent leading to the increase in excited electrons on the catalyst surface for hydrogen production from protons.¹⁹ In an experiment, Nada et.al., compared several sacrificial agents for H₂ production and listed their findings in the sequence: EDTA > methanol > ethanol > lactic acid.³⁴ Based on this reactivity series, Chen et al. documented that the structures of the sacrificial reagents affect the rate of hydrogen production and it can be enhanced by the presence of α-H adjacent to the OH groups.³⁵ Police et al. reported that the glycerol with 5 α-H atoms produces greater hydrogen as compared to ethanol with 2 α-H atoms.³⁶ Ni et al. and Abe et al. examined the effect of inorganic ions such as $IO3−/I−$⁠, which act as sacrificial reagents for photocatalytic H₂ production. The results of their experiments show an enhancement in H₂ production as the I⁻ ions (electron donor) fill the generated holes in VB and, thus, lead to the availability of CB excited electrons for hydrogen production.^37,38

Doping with cationic metals (transition and noble metals) at Ti sites (e.g., Au, Pt, Ag, Cr, V, Fe, and Ni) and anionic non-metals (e.g., N, S, and C) at O sites is an important strategy for attuning the bandgap (to some extent) and catalytic properties of TiO₂ for water splitting. Khairy and Zakaria found that dopant can enhance (e⁻/h⁺) migration and separation. Modified titania with dopants cause shifting of optical absorption to longer wavelength as it offers extra active site locations on the surface of TiO₂, thus, demanding minimum overpotential for hydrogen production.³⁹ 

Metal loading on the surface of TiO₂ has become the most effective route to enhance the electron–hole separation.⁴⁰ Doping with metals help to generate donor energy levels over the valance band of TiO₂ or accepter levels beneath the conducting bands of TiO₂ (Fig. 2). This helps in lowering the bandgap energy and, consequently, enhances the activity of photocatalysts. The metal can act as the electron sink in which the photoelectrons move easily from the TiO₂ to the metal via the Schottky barrier.

Recently, several noble and transition metal oxides of Fe, Cu, Pt, Au, Ag, and Pd have been investigated. Table II summarizes the amount of hydrogen production resulting from metal dopants. A comparison of the hydrogen generation efficiency in the absence and presence of catalyst clearly demonstrates the role of catalyst in enhancing water splitting. Among different metallic dopants, noble metals (e.g., Pt, Au, and Ag) are very effective toward H₂ production as they extend the photocatalytic performance to the visible region due to their strong surface plasmon bands.³³ As a result, the photoelectron transfers from metal to the CB of TiO₂ causing effective reduction reaction at the surface of the semiconductor. Figures 3(a) and 3(b) illustrate the schematics of the mechanism of the metal/TiO₂ absorption.

Like the noble metals, doping with transition metals is deemed as an effective approach to increase the performance of titania. In this regard Ni et al., documented that the different effects of metal ions toward the activity of photocatalyst relate to their different mechanisms of electrons/holes migration.³⁷ Among a variety of metal doping candidates for TiO₂, Allen et al. reported Cu as an effective co-catalyst for photocatalysis. Owing to its high abundance in the Earth’s crust, Cu is 100 times more cheaper than noble metals.⁴⁸ Moreover, Zuas and Budiman reported that cupric ions can trap e⁻ and h⁺ and form extra energy levels close to TiO₂’s VB and CB.⁴⁹ 

Recently, transition metal nitrides have also been used as plasmonic materials for photocatalytic processes as the optical properties of transition metals are almost the same as those of the noble metals.⁵⁰ However, the plasmon resonance and chemical stability of transition metals are higher than the noble metals; hence, they cause to transfer more electrons into the conduction band of titania as compared to noble metals. Transition metal oxides are categorized into two groups. Early transition metals have empty d orbitals, e.g., Ti, Sc, and Nb. Therefore, their valence band is strongly affected by 2p oxygen orbitals. As a consequence, there are wide bandgaps in these metals and they show low photocatalytic activity as compared to late transition metals that possess small bandgaps and occupied d orbitals (e.g., Mn, Fe, Co, and Ni). Jaafarzadeh et al. reported that among the doping transition metal ions, Fe³⁺ is considered as a good dopant due to its half-filled 5d orbitals.⁵¹ Whereas, Ni₂O₃ and CuO are considered promising photocatalysts due to their plasmon absorption properties under visible light radiation.⁵² 

Light irradiation also influences photocatalytic hydrogen generation. In this regard, Zhang et al. documented that electron transfer is strongly dependent on light irradiation; for instance, Au–TiO₂ follow two opposite processes: (i) under UV light (TiO₂ to Au electronic transition) and (ii) under visible light irradiation (Au to TiO₂ electronic transition).⁵³ Similarly, Liu et al. found a greater hydrogen production rate from UV-Visible irradiation as compared to the sum of H₂ produced from separate radiations.⁵⁴ The enhancement can be related to the synergistic effect of Schottky barrier formation and surface plasmon resonance.

Doping with anionic non-metals is another option of lowering the bandgaps of semiconductors. In this regard Chen et al. reported that doping with non-metals produces hybrid 2p level between the conduction and valence bands of the TiO₂.⁵⁵ Radiations in the visible range cause to shift electrons from VB of TiO₂ to the intermediate 2p band and, then, excite to the CB for producing hydrogen. In addition, Luo et al. mentioned that non-metal dopants are more effective and approachable toward increasing the visible light absorption than metal dopants.⁵⁶ A prominent decrease in the bandgaps has been observed by doping TiO₂ with a variety of anion dopants such as anions of N, S, etc. Chen and Burda used x-ray photoelectron spectroscopy to study the electronic effect of C⁻, N⁻, and S⁻ doped TiO₂, and the results showed additional energy levels on top of the VB of TiO₂.⁵⁷ Moreover, Wang et al. found enhancement in the performance of titania doped with N and got higher hydrogen production rate on doped titania as compared to un-doped TiO₂. Thus, substitutional N is considered to be the dominating factor of increasing photocatalytic property of titania.⁵⁸ Similarly, Xing et al. reported the activity of S-doped titania and found that S atoms lower the bandgap of doped titania and enhance the utilization of the visible light radiation.⁵⁹ In addition, Luo et al. doped bromide and chloride ions in TiO₂ via the hydrothermal process and the resulting photocatalyst was found to lower the bandgap and, thus, extend absorption to the visible region.⁵⁶ Moreover, Wang et al. demonstrated that doping with carbon nanotubes leads to enhancement of the photocatalytic activity due to the synergetic effect of carbon on TiO₂.⁶⁰ Three mechanisms were proposed for the synergetic effect: the first one proposes carbon as an electron sink, which results in preventing the electron–hole pair recombination process.⁶¹ The second mechanism considers carbon to act as a photosensitizer causing the pumping of the photoexcited electrons to the conducting band.⁶² The third mechanism assumes that carbon hinders the agglomeration of TiO₂ particles. The quality of semiconductors can be improved by applying the donor–acceptor co-doping approach. Dong et al. reported that (Ti⁺³, Ni) co-doped TiO₂ is an effective nonmetal–metal candidate for modifying the activity of catalyst by narrowing the bandgap to about 2.84 eV, which attributes to improve the visible light absorption.⁴⁴ 

Although, doping of metals and/or nonmetals into photocatalysts is one of the efficient approaches for extending light absorption, the effectiveness may be compromised because of dopants. They may disturb the photocatalyst surface by unsuccessfully moving the photoexcited charge or the dopants sites, which may act as recombination sites. Recent studies have proved that coupling TiO₂ with binary composite and transition metal oxides such as SiO₂, Al₂O₃, CdS, Cu₂O, Fe₂O₃, and ZnO can improve the catalytic activity as they decrease the bandgap and increase the visible light utilization. In this connection, Table III summarizes the photocatalytic hydrogen production that results from semiconductors coupled with TiO_2. The photocatalytic performance is clearly evidenced from the significantly higher generation of hydrogen from the aqueous system in the presence of catalysts as compared to water splitting without photocatalysts.

The phenomenon based on transferring the electrons from the semiconductor with low bandgap to the larger one is illustrated in Fig. 4. Wu et al. and Huy et al. studied the coupling of TiO₂ and SrTiO₃ and reported that the electrons transferred from CB of SrTiO₃ to TiO₂, thus, preventing the recombination of e⁻/h⁺ pairs and improving the photocatalytic activity.^67,68 Moreover, Huy et al. reported that the appropriate morphology of the catalyst can modify the performance of the metal by increasing its surface area.⁶⁸ Similarly enhancement in the photocatalytic performance of the nanogranular structure of SrTiO₃/TiO₂ was reported by Huy et al.⁶⁸ Moreover, Wu et al. investigated the photocatalytic water splitting enhancement in the BiFeO₃/TiO₂ composite (BFO/TiO₂).⁶⁷ The results of their findings show that the photocurrent density reaches 11.25 mA/cm², which is more than 20 times than that of bare TiO₂.⁶⁹ The UV-visible spectra of BFO/TiO₂ showed a sharp edge of the absorption band for bare TiO₂ (emerged below 400 nm); however, visible light utilization was found to significantly increase after depositing the BFO thin films.

Besides coupling with narrow bandgap semiconductors, researchers investigated the impacts of coupling TiO₂ with wide bandgap semiconductors. In one particular study, coupling TiO₂ with SiC₂ (bandgap = 3.0 eV) was found to generate hydrogen efficiently under UV irradiation due to efficient charge separation caused by transferring the electrons from the CB of TiO₂ to SiC₂.⁷⁰ Working in this area, Nguyen et al. argued the ability of two photocatalytic systems, TiO₂–SiO₂ and RuS₂/TiO₂–SiO₂ toward water splitting.⁷¹ They noticed that the TiO₂–SiO₂ semiconductor carries more negative CB compared to TiO₂. However, further combination of TiO₂–SiO₂ with RuS₂ allows more electrons to be transferred to the CB of RuS₂, leading to more hydrogen reduction.

Recently, the composite of TiO₂ and carbon nanomaterials, especially carbon nanotubes and graphene have attracted much attention. Bashiri et al., commented that the introduction of the graphene on the TiO₂ surface expands the utilization of the visible region, facilitates charge separation, and increases the number of reaction sites.⁶ Moreover, the studies found that the ternary photocatalysts (three semiconductors or elements) could increase the performance of TiO₂ toward hydrogen production. Multi-photonic excitation with reduced energy photon could be anticipated from ternary semiconductor composites. Ternary photocatalysts enhance the production of H₂ by improving light harvesting and mass transfer initiation of TiO₂–Au–CdS.

In order to replace the fossil fuels, many attempts have been made to develop effective and renewable energy fuels. Water splitting plays an important role in producing clean energy in the form of H₂. The selection of a suitable photocatalyst and adoption of an effective strategy are of primary importance for achieving a higher production rate of hydrogen. Many catalysts have been tested; the challenge, however, lies in comparing different materials to achieve practical implementation for H₂ production. TiO₂ is considered as a promising semiconductor, but its performance is still low, primarily due to its low visible light utilization. Researchers have found that the ionic co-catalyst (as a dopant) and other semiconductors (as a heterojunction composite) have a major impact on the photocatalytic performance of TiO₂ by reducing the bandgap and controlling the morphology of TiO₂. Future research in catalyst innovations and surface modification is essential to adjust the bandgap and improve visible light absorption. The probability of modifying TiO₂ for more effective water splitting depends on studying all the operating parameters such as temperature, pH, and reactor type. Therefore, both theoretical and experimental studies are recommended for understanding the principle of water splitting using modified TiO_2.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

The authors declare no conflict of interest.