**4.2 Solar fuel production**

In recent years, semiconductor-based materials have been extensively studied for energy applications that can contribute to reducing greenhouse gases. Storing solar energy into the chemical bonds of fuel seems to be a promising way to replace the traditional combustion of fossil fuels with environmentally friendly technology. Specifically, much attention has been paid to photocatalytic H2 production from water and CO2 reduction to valuable chemical compounds.

Regarding H2 generation from water, it should be noted that pristine structures show low activity, and therefore, surface modification with co-catalysts needs to be applied. In the case of the faceted particles, the overall problem deals with specific interactions between the surface and co-catalyst, as well as possible charge separation between different co-exposed facets. Focusing on the single-facet, important findings were reported by Gordon et al., who noticed the higher activity of the anatase modified with Pt for octahedrons exposing {1 0 1} than nanosheets with exposed {0 0 1}

[67]. Similar results were obtained by Wang et al. for anatase nanoparticles modified with MoXC as a co-catalyst [81]. This fact is attributed to the increased reduction ability of the (1 0 1) surface, which results in the synergy of TiO2 and co-catalyst. Recently, specifically, the problem of interactions between (1 0 1) and different possible metal co-catalyst was investigated in detail by Wang and Gong in their computational study [82]. Based on the obtained results, they have proposed alloyed Cu/Pt and Rh/Pt co-catalysts as promising candidates for hydrogen evolution. This concept was based on optimizing the electron transfer between (1 0 1) anatase surface and Cu or Rh as the electron-acceptor and further exposition of Pt as the active part of the co-catalyst. Considered models and their electronic structures are shown in **Figure 10**.

Furthermore, a combination of {1 0 1} with other co-exposed facets can increase the activity of the TiO2 materials for H2 generation. For example, Wei et al. presented a detailed comparison between octahedral {1 0 1} and decahedral {1 0 1}/{0 0 1} anatase particles modified with Cu, Ag, and Au nanoparticles. Particularly, a combination of both {1 0 1} and {0 0 1} facets resulted in a slightly higher activity when modified with Ag or Au, as well as a significantly higher activity when modified with Cu [83]. Furthermore, Meng et al. reported increased H2 production using the decahedral {1 0 1}/{0 0 1} anatase samples, when both facets were selectively modified by Pt and Co3O4, respectively [84]. It is especially noteworthy that such a combination of the facet co-exposition and selective modification with optimized cocatalysts was recently proposed to achieve almost 100% of quantum efficiency during water splitting reaction over SrTiO3 photocatalyst [4]. Therefore, it confirms the

### **Figure 10.**

*Optimal models (a, c) for electron transfer between TiO2 (1 0 1) anatase surface and alloyed metal cocatalysts, proposed by Wang and Gong and their corresponding density of states distribution (b, d). In images (b, d), blue line shows the states of the metal cluster, while yellow-green is TiO2. Reprinted from the [92] under a creative commons attribution 4.0 international license. IET refers to the energy of intrinsic electron transfer in eV.*

importance of optimizing facet-facet exposition and further facet-co-catalyst interactions to optimize the final performance.

Furthermore, recent studies also focus on the photocatalytic reduction of carbon (IV) oxide to valuable chemical compounds. This reaction begins with the adsorption of CO2 and H2O molecules, which was investigated theoretically by Mishra and Nanda. Using DFT calculations, they examined the chemical restructuring of CO2 and H2O molecules during the process of adsorption, co-adsorption, and conversion on (0 0 1), (1 0 0), and (1 0 1) surfaces [85]. They observed that the energy barrier of bicarbonate complex formation, which resulted from the co-adsorption of carbon dioxide and water, was the lowest for the (0 0 1) surface. Therefore, {0 0 1} facets are supposed to be the most reactive anatase facets for CO2 photocatalytic reduction. However, if this surface undergoes reconstruction, the number of active sites is reduced. Therefore, experimental conditions like temperature and high vacuum will be crucial for the photocatalytic performance of anatase nanocrystals.

Although the photocatalytic reaction depends on the adsorption of reactants, the investigations provided by Ma et al., in the application of CO2 reduction to formic acid, showed different behavior of anatase crystal facets compared with previous studies [86]. The surface electron transfer for (0 0 1) and (1 0 1) surfaces was characterized by similar barrier levels. However, the reductive ability of electrons generated on the (1 0 1) plane is higher than that on the (0 0 1) plane; therefore, electrons may be transferred more easily to reactants for low-energetic facets. Moreover, HCOOH on the (0 0 1) surface can replace water and, in consequence, occupy the active sites, hindering the reaction. On the contrary, formic acid seemed to remain undissociated on (1 0 1) surfaces, so more suitable product adsorption properties led to a higher photocatalytic performance.

The above-reported studies were theoretical, so the experimental results may not be consistent with DFT calculations. Therefore, Liu et al. demonstrated the blue anatase nanocrystals with exposed {0 0 1}, {1 0 1}, and a combination of {1 0 1} and {0 0 1} facets [87]. They reported that oxygen-deficient TiO2 nanostructures with co-exposed {1 0 1} and {0 0 1} facets exhibited relatively high quantum yield for CO2 reduction to CO (0.31% under UV–vis light and 0.134% under visible light). Moreover, this photocatalyst demonstrated more than four times higher visible light activity in comparison with {0 0 1} or {1 0 1}. This high photocatalytic activity was a result of two effects. Firstly, co-exposed {0 0 1} and {1 0 1} facets had increased the capacity of reversible CO2 adsorption. Secondly, the created surface junction between facets enhanced the charge separation and hindered the recombination processes. Similar results were obtained by Yu et al., who investigated the mist-efficient content ratio of {0 0 1} and {1 0 1} facets [88]. The decahedral-shaped sample with 58% content of {0 0 1} facets exhibited the highest methane production from CO2. {1 0 1} facets acted as reduction sites, whereas {0 0 1} facets were the oxidation sites on the photocatalyst surface. However, a too high amount of the {0 0 1} facets on the anatase surface may have caused an electron overflow effect toward {1 0 1} facets, so the migration of electrons to {1 0 1} facets is more difficult than in the previous case.

Carbon dioxide may be further converted to methane, which generally gives rise to operational risks and environmental problems [89, 90]. Therefore, selective oxidation to CH3OH is a promising way to CH4 storage. Feng et al. reported the facet-dependent selectivity of CH4 ! CH3OH conversion over anatase nanocrystals. They showed that silver-decorated TiO2 with predominant {0 0 1} facets exhibited a selectivity of approximately 80%, which was significantly better than the sample with dominated {1 0 1} facets. This high selectivity resulted from oxygen vacancy generation by

### **Figure 11.**

*The proposed mechanism of methane photocatalytic conversion over {0 0 1} anatase crystal facet. Reprinted from [101] under a creative commons attribution 4.0 international license.*

photoinduced holes, which played a crucial role in avoiding the formation of CH3 and OH radicals. Therefore, the undesired overoxidation to CO was limited, in opposite to TiO2 exposing {1 0 1} facets [91]. The proposed mechanism of CH4 oxidation on the {0 0 1} facets is presented in **Figure 11**.
