Date Published: March 12, 2018
Publisher: John Wiley and Sons Inc.
Author(s): Yuchan Dong, Kulbir Kaur Ghuman, Radian Popescu, Paul N. Duchesne, Wenjie Zhou, Joel Y. Y. Loh, Abdinoor A. Jelle, Jia Jia, Di Wang, Xiaoke Mu, Christian Kübel, Lu Wang, Le He, Mireille Ghoussoub, Qiang Wang, Thomas E. Wood, Laura M. Reyes, Peng Zhang, Nazir P. Kherani, Chandra Veer Singh, Geoffrey A. Ozin.
Frustrated Lewis pairs (FLPs) created by sterically hindered Lewis acids and Lewis bases have shown their capacity for capturing and reacting with a variety of small molecules, including H2 and CO2, and thereby creating a new strategy for CO2 reduction. Here, the photocatalytic CO2 reduction behavior of defect‐laden indium oxide (In2O3−x(OH)y) is greatly enhanced through isomorphous substitution of In3+ with Bi3+, providing fundamental insights into the catalytically active surface FLPs (i.e., In—OH···In) and the experimentally observed “volcano” relationship between the CO production rate and Bi3+ substitution level. According to density functional theory calculations at the optimal Bi3+ substitution level, the 6s2 electron pair of Bi3+ hybridizes with the oxygen in the neighboring In—OH Lewis base site, leading to mildly increased Lewis basicity without influencing the Lewis acidity of the nearby In Lewis acid site. Meanwhile, Bi3+ can act as an extra acid site, serving to maximize the heterolytic splitting of reactant H2, and results in a more hydridic hydride for more efficient CO2 reduction. This study demonstrates that isomorphous substitution can effectively optimize the reactivity of surface catalytic active sites in addition to influencing optoelectronic properties, affording a better understanding of the photocatalytic CO2 reduction mechanism.
Photocatalytic reduction of carbon dioxide (CO2) has been explored for many years as an attractive strategy for reducing CO2 emissions while producing renewable fuels and chemicals.1, 2, 3 Oxide semiconductors are widely used as CO2 reduction photocatalysts due to their light‐harvesting property, high stability, and low cost.4, 5, 6, 7 However, due to the high thermodynamic stability of CO2 and the limitations of intrinsic semiconductor photocatalysts, such as poor charge separation and weak interaction with reactants, intensive efforts have been made to improve these oxide semiconductor photocatalysts via doping.8, 9 For instance, in fluorinated anatase TiO2 nanosheets prepared via substitution of surface hydroxyl groups with fluoride anions, surface fluorination promoted the formation of Ti3+ defects that helped extend the lifetime of photogenerated electrons and holes, and facilitated the reduction of CO2 to CO2−.10 In another study, copper‐loaded TiO2 catalysts prepared via a sol‐gel method showed a great increase in methanol production rate relative to nondoped TiO2. In this case, Cu+ also inhibits the recombination of photoexcited charge carriers without affecting electron mobility.11 Despite the large number of doped oxide semiconductor photocatalyst studies, the focus has been mostly on photoelectrochemical CO2 reduction, and explanations for the improved catalytic activity are still mainly limited to optoelectronic properties such as bandgap, light absorption efficiency, charge carrier lifetimes, and electron mobility.12, 13, 14, 15 However, like other catalytic reactions, the efficiency of photocatalytic CO2 reduction relies heavily on the nature of catalytic sites at the photocatalyst surface, where the adsorption, activation, and reaction of gaseous reactants, and desorption of products, occur.16, 17, 18, 19, 20 Thus, it is important to examine how doping affects the surface chemistry and catalytic process in order to rationally design catalytic materials capable of replacing nonrenewable fossil fuels with renewable synthetic fuels through photocatalytic CO2 reduction.
In pursuing heterogeneous catalysts capable of facilitating the efficient conversion of gaseous CO2 to useful chemicals and fuels, it has been demonstrated that the nature of surface FLP sites can serve as an effective descriptor for identifying the best catalyst in an isostructural series of materials. This strategy is exemplified herein by the archetypical material BizIn2−zO3−x(OH)y, which catalyzes the photochemical RWGS (CO2 + H2 → CO + H2O). By varying the concentration of Bi3+ in the lattice of nanocrystalline In2O3−x(OH)y, a maximum in the “volcano” plot for the conversion of CO2‐to‐CO was found to occur at z = 0.0003. Density functional theory calculations suggested that the uphill portion of the “volcano” curve can be traced to the increase in both Lewis acidity and Lewis basicity of the FLP when Bi substitutes at the neighboring In3 site. Initially, this serves to improve the heterolytic splitting of reactant H2, whereupon the rate of the RWGS reaction increases. When the Bi‐substitution was further increased, acidic In2 atoms could also have been substituted by Bi, leading to a decrease in reaction rate due to a reduction in the FLP site strength, thereby giving rise to the downhill portion of the “volcano” plot. This is consistent with the experimental observation that the CO2 reduction rate decreased rapidly at higher Bi concentrations (i.e., 0.5%), even though the number of oxygen vacancy and surface hydroxide defects was high and the optoelectronic properties were enhanced. Based on the results of this study, it is anticipated that the strength of surface FLP site, which can be effectively tuned through isomorphous substitution, will become an increasingly valuable parameter in the search for CO2 reduction materials exhibiting superior catalytic performance.
Preparation of In2O3−x(OH)y: In2O3−x(OH)y nanocrystals, denoted as “Bi‐0%,” were prepared via dehydroxylation of In(OH)3 in air at 250 °C for 6 h. In a typical synthesis of In(OH)3, indium(III) nitrate hydrate (1.8 mmol, 0.54 g) was dissolved in ethanol (6 mL), to which was then added an ammonia solution prepared by mixing ammonium hydroxide (27%, 2.5 mL) with ethanol (7.5 mL) and water (2 mL). The resulting suspension of In(OH)3 was kept in a preheated oil bath at 80 °C for 10 min and then cooled to room temperature before the precipitate was collected via centrifugation and washed with deionized water. The resulting solid was dried at 60 °C in a vacuum oven.
The authors declare no conflict of interest.