**1. Introduction**

Since the discovery of water splitting on a TiO2 electrode under irradiation of ultraviolet (UV) light by Fujishima and Honda in 1972, many semiconductors, such as CdS, ZnO, SrTiO3 and g-C3N4, have been demonstrated to exhibit photocatalytic properties and, through the efforts of researchers, the application areas of semiconductor photocatalysts have been greatly extended [1–14]. Unfortunately, due to fast charge separation and limited light absorption, the properties of semiconductor photocatalysts are relatively unsatisfactory for practical application [15–17]. To improve the properties of semiconductor photocatalysts, several strategies have been proposed by researchers, such as doping, dye-sensitization, coupling, etc. [18–27]. However, each of these strategies has its own pros and cons. Toward the strategy of doping, the band structures of semiconductor photocatalysts could be modulated by the doping atoms to exhibit better properties for light absorption, but the doped semiconductor photocatalysts could be more susceptible to photo-corrosion and the charge recombination of the doped semiconductor photocatalysts could be aggravated at the defects introduced by the doping atoms [19,21,27]. As for the strategy of dye-sensitization, although the light-sensitive dyes can absorb light more efficiently and transfer the photoexcited electrons to the semiconductor photocatalysts, the light-sensitive dyes are susceptible to chemical corrosion, resulting in the poor stability of dye-sensitized semiconductor photocatalysts [24–26]. Coupling, as a strategy with a relatively short history, has aroused grea<sup>t</sup> interest among researchers since its very beginning. When coupled with metals, especially the noble metals, the properties of semiconductor photocatalysts can be considerably improved due to the enhanced charge separation at the metal–semiconductor interface and the enhanced visible light absorption caused by the surface plasmon resonance (SPR) of metals [28–30]. Except for the high cost of noble metals, the main drawback of the coupling strategy used to be the poor control of the process of coupling semiconductors with metals. However, with the development of nanotechnology, the coupling process now could be delicately controlled and several new structures have been synthesized to improve the properties of metal/semiconductor (M/S) nanocomposite photocatalysts [29,31].

Herein, we would like to review the M/S nanocomposite photocatalysts regarding their fundamentals, structures (as well as their typical synthetic approaches), applications and properties. Finally, we will also present our perspective on the future development of M/S nanocomposite photocatalysts.
