*3.2. Heterojunctions and Photocatalysis*

Electron–hole charge pairs formed by the photocatalytic hydrogen evolution reaction are transferred to the surface of the photocatalyst or else recombine with each other. To better understand this point, let us illustrate it by reviewing the presentation in [63]: a comparison of the influence of gravitational force on a man jumping off the ground and electrons jumping from the VB to the CB (Figure 6a,b, respectively). If a man (electron) jumps from the ground (VB) into the sky (CB), it will return to the floor quickly (recombine with the hole) due to gravitational force. However, a stool (semiconductor B) can be provided to get the man off the ground (separate the photogenerated electron–hole pair), as illustrated in Figure 6c,d, respectively. Subsequently, the aforementioned man will land again on the stool rather than the ground (the electron–hole pair recombination will be inhibited). Preventing electron–hole recombination is an urgent issue, but it can be achieved by the proper design of materials [64–66]. Many methods have been proposed to achieve better separation of the photogenerated electron–hole pairs in semiconductor photocatalysts, such as element combining, metal and non-metal doping, and heterojunctions [67–72]. Among these strategies, heterojunctions in photocatalysts have proved to be one of the most promising methods for efficient photocatalyst preparation due to their improved separation of electron–hole pairs [73].

**Figure 6.** Schematics of (**a**) the influence of gravitational force on a man jumping; (**b**) recombination of photocatalyst electron–hole pair; (**c**) using stool to keep a man from returning to the ground; and (**d**) electron–hole pairs separated in a heterojunction photocatalyst. Reproduced with permission from [63]; copyright (2017), John Wiley & Sons, Inc.

#### 3.2.1. Semiconductor Heterojunction Photocatalysts

Suppressing the electron–hole recombination rate is the most important solution to increase photocatalytic efficiency. Bulk g-C3N4 has low ability to collect visible light, low charge-transport properties, and small surface area, so there have been many studies to make it an efficient photocatalyst [74]. Various strategies have been proposed to achieve better electron–hole pair separation such as element combining, metal doping, and creating heterojunctions. Among these strategies, g-C3N4/semiconductor heterojunctions have shown the improved separation capability of electron–hole pairs; the charge carrier is transferred through the heterostructure interface to inhibit recombination, thereby improving the photocatalytic performance [75–77]. In addition, a g-C3N4/semiconductor heterostructure can be formed by combining a visible-light excited photocatalyst semiconductor material having a narrow band-gap and a photoexcited photocatalyst having a large band-gap in a coupling process; the connection between the two different kinds of photocatalyst having different band structures induces a new band arrangement [78,79].

In 2017, Zhang et al. [80] reported the in situ synthesis of a g-C3N4/TiO2 heterostructure photocatalyst which greatly improved the hydrogen evolution performance under visible light. The g-C3N4 nanosheets were synthesized by calcining urea at 550 ◦C for 4 h. Two hundred milligrams of the as-prepared g-C3N4 nanosheets were dispersed in 20 mL ethanol and sonicated for 1 hour to obtain a homogeneous suspension. Under continuous stirring, 40 mL of ammonia solution (~28 wt%) and tetrabutyl titanate (TBT) (0, 100, 200, 300 and 400 μL) were added and stirred for 12 h to achieve the in situ synthesis of amorphous TiO2. The obtained products were expressed as CNTO-*x* (*x* = 0–4) according to the TBT content. As shown in Figure 7a, the shape of the CNTO-2 sample seen in a transmission electron microscopy (TEM) image shows that the TiO2 nanoparticles are uniformly distributed in the g-C3N4 nanosheets. As a result, there is uniform interfacial contact between the

TiO2 phase and the g-C3N4 phase. Figure 7b shows the average rate of hydrogen production within 3 h. Pure TiO2 does not react with visible light and produces negligible H2, while CNTO-0 exhibits a low hydrogen production rate of 15 μmol h−<sup>1</sup> due to the fast recombination of photogenerated charge carriers. In contrast, the CNTO-2 sample exhibits significantly improved hydrogen production performance at 40 μmol h−1. However, as the amount of TiO2 is further increased, TiO2 occupies the surface of g-C3N4 resulting in less active sites for H2 evolution. The proposed mechanism of heterostructure composites is also shown in Figure 7c. According to previous reports, the CB and VB potentials of g-C3N4 and TiO2 are −1.12 and +1.58 V, and −0.29 and +2.91 V, respectively. Under visible light irradiation, only g-C3N4 can absorb light to generate electron–hole pairs. However, in pure g-C3N4, photogenerated electrons and holes recombine rapidly, and only a few of the electrons participate in the reaction, resulting in low reactivity. When g-C3N4 is modified by TiO2 to form a heterojunction structure, the CB edge of g-C3N4 is more negative than TiO2, so that electrons excited in the CB of g-C3N4 can be injected directly into the CB of TiO2. Consequently, Pt2<sup>+</sup> adsorbed on the surface is reduced by electrons transferred from the CB of TiO2, and newly formed Pt nanoparticles are deposited on the surface of TiO2 as an efficient cocatalyst for hydrogen production. The electrons then accumulate in Pt nanoparticles and participate in hydrogen evolution. Therefore, the photocatalytic activity of the g-C3N4/TiO2 composite with Pt nanoparticles as a cocatalyst is significantly improved.

**Figure 7.** (**a**) A TEM image of CNTO-2, (**b**) H2 evolution rates of the CNTO-*x* samples under visible light (λ ≥ 420 nm), and (**c**) an illustration of the g-C3N4/TiO2 heterojunction system. Reproduced with permission from [80]; copyright (2017), The Royal Society of Chemistry.
