Toward Enhanced Photocatalytic Oxygen Evolution: Synergetic Utilization of Plasmonic Effect and Schottky Junction via Interfacing Facet Selection

DOI: 10.1002/adma.201501200 participate in oxidation reaction. As long as a semiconductor with appropriate bandgap (i.e., wide bandgap) is selected, the redox abilities of electrons or holes can be maintained as high as those in wide-bandgap semiconductors despite the use of incident visible light with relatively low energy. Unfortunately, the reported photocatalytic effi ciencies purely offered by the plasmonic hot carrier injection effect in the absence of semiconductor photoexcitation are negligible in contrast to those by semiconductor photoexcitation. [ 11,19 ] The major reason for this limitation is the lack of a driving force to steer the migration of injected electrons or holes to semiconductor surface for reduction or oxidation reactions. The low charge migration rates and uncertain charge diffusion directions make the charge carriers randomly walk in the semiconductor, so only a small portion of plasmonic hot carriers can arrive at the catalytic sites. We have thus decided to develop a new approach to better harness the utilization of plasmonic hot carriers. Thus far, use of a Schottky junction has been recognized as the most wellestablished strategy for steering the fl ow of the carriers that are photogenerated in semiconductor. It is well known that metal (especially for nonplasmonic metal, Pt and Pd) can serve as an sink for the photogenerated electrons or holes when forming a Schottky junction with n-type or p-type semiconductor, respectively (Figure S1, Supporting Information). [ 20,21 ] The formed Schottky barrier can inhibit the backfl ow of electrons or holes from metal to semiconductor. As a result, the charge “pump” role of the Schottky junction ensures the effi cient unidirectional transfer of charge carriers across the interface of metal– semiconductor (M–S) junction. Naturally we consider the possibility whether this Schottky-junction effect may be extended to the utilization of plasmonic hot carriers in photocatalysis through guiding their migration directions. However, this idea can be hardly accomplished by a single M–S junction between plasmonic metal and semiconductor. When a plasmonic metal is used for both the Schottky junction and hot carrier injection, the injection of plasmonic hot carriers would follow an opposite direction to the carrier fl ow driven by the Schottky junction (Figure S1, Supporting Information). [ 11,14,15,18,22 ] This severe competition dramatically reduces the effi ciency of carrier trapping on metal and e–h separation, particularly when metal and semiconductor are both excited under full-spectrum irradiation. In this communication, we report a new design for synergizing the plasmonic effect with the Schottky junction. The core concept of this work is to separate the Schottky junction from the plasmonic hot carrier injection by building two M–S interfaces based on the selection of semiconductor facets and metals. The functions of these two interfaces are synergized by Photocatalytic water splitting represents a highly important approach to addressing current energy and environmental demands. Photocatalysis requires effi cient separation of photo generated electron–hole (e–h) pairs in semiconductor to undergo redox reactions. [ 1 ] The reduction and oxidation capabilities of photogenerated electrons and holes in a semiconductor are determined by the positions of conduction band (CB) and valence band (VB) edges, respectively. Only when the CB edge lies at a higher position (more negative) than the redox potential of reduction half reaction, and meanwhile, the VB edge is at a lower position (more positive) than the potential of oxidation half reaction, can an overall photocatalytic reaction take place. [ 2 ] Thus wide-bandgap semiconductors with higher CB and lower VB edges generally show higher redox abilities as well as more promising photocatalytic performance in comparison with narrow-bandgap ones. However, semiconductors with wide bandgaps can only absorb light in the UV region which accounts for ≈5% of solar spectrum, thereby limiting their solar energy conversion effi ciency for practical applications. [ 3 ] For this reason, the relationship between absorption of long-wavelength light and high redox abilities of charge carriers is essentially an irreconcilable contradiction for bare semiconductor photocatalysts. Most recently, integration of surface plasmon into photocatalysis has been widely explored by composing hybrid structures between noble metals and semiconductors, which may potentially circumvent this situation. [ 4–10 ] As demonstrated by many research groups, [ 11–18 ] the metal with surface plasmon (e.g., Ag and Au) that directly contacts a semiconductor can be excited under visible light illumination to generate and inject hot carriers into the semiconductor. Specifi cally, hot electrons may fl ow into the CB of n-type semiconductor [ 13 ] and in the case of p-type semiconductor, instead hot holes are injected into the VB of semiconductor (Figure S1, Supporting Information). [ 16 ]