Article pubs.acs.org/JPCC
Redox Dynamics of Rh Supported on ZrP2O7 and ZrO2 Analyzed by Time-Resolved In Situ Optical Spectroscopy Haris Puspito Buwono,†,§ Masahiro Yamamoto,† Riichiro Kakei,† Satoshi Hinokuma,†,‡ Hiroshi Yoshida,†,‡ and Masato Machida*,†,‡ †
Department of Applied Chemistry and Biochemistry, Graduate School of Science and Technology, Kumamoto University, 2-39-1 Kurokami, Chuo, Kumamoto 860-8555, Japan ‡ Unit of Elements Strategy Initiative for Catalysts & Batteries, Kyoto University, 1-30 Goryo-Ohara, Nishikyo, Kyoto 615-8545, Japan § Department of Mechanical Engineering, State Polytechnic of Malang, Jl. Soekarno Hatta No. 9, Malang 65141, Indonesia S Supporting Information *
ABSTRACT: In situ time-resolved diffuse reflectance spectroscopy was first applied to supported Rh catalysts (0.4 wt % Rh/ZrO2 and Rh/ZrP2O7) under dynamic three-way catalysis conditions fluctuating between fuel-lean and fuel-rich gas atmospheres. The optical absorption at 650 nm was found to decrease upon lean-to-rich switching of the gas feed, which led to the reduction of Rh oxide (Rh3+) to metallic Rh (Rh0), followed by a reversible increase upon back switching rich-to-lean. The kinetic analysis suggested that the reduction of Rh3+ to Rh0 was faster than the reoxidation over Rh/ ZrP2O7, whereas the reduction was comparable with or slower than the reoxidation over Rh/ZrO2. The activation energy of Rh/ZrP2O7 for the reduction, 13.6 kJ mol−1, was smaller than that for the oxidation, 48.7 kJ mol−1, which contrasted with those of Rh/ZrO2 (21.4 and 34.1 kJ mol−1, respectively). These results were closely associated with the higher NO reduction activity of Rh/ZrP2O7 than Rh/ZrO2 under a lean-gas atmosphere because Rh was more active in the metallic state than in the oxide state. Applying fast lean−rich perturbation of the gas feed with 1 s intervals led to an immediate and significant drop of the optical absorption intensity, suggesting that the reduction of Rh substantially penetrated to deeper layers under the surface. This study provided the first in situ evidence for the formation of active metallic Rh species under high-frequency lean−rich oscillations.
1. INTRODUCTION Rhodium is well-known as an active metal component for NO reduction in three-way catalysis (TWC), and the activity of Rh changes depending on the oxidation state.1−6 In TWC, the catalyst is exposed to a dynamic oxidation−reduction perturbation atmosphere, which is expressed in terms of an air-to-fuel ratio (A/F) on a weight basis. A reducing (fuel-rich) condition occurs below the stoichiometric ratio (A/F < 14.6), whereas an oxidizing (fuel-lean) condition occurs at A/F > 14.6. The metallic Rh (Rh0) formed during A/F < 14.6 exhibits high NO conversion, whereas the oxide (Rh3+) formed during A/F > 14.6 is less active. We recently reported that Rh supported on metal phosphates showed a unique characteristic: the Rh oxide was more easily reduced to the active metallic Rh, which was only slowly oxidized compared with Rh supported on metal oxides.7−10 One practical example of importance was that a ZrP2O7 support significantly enhanced the NO conversion efficiency during A/F > 14.6 compared with the ZrO2 support.8,10 The efficiency was further enhanced by switching the two simulated gas feeds (stoichiometric (A/F = 14.6) and lean (A/F = 14.9)) at short intervals (∼1 s).10 Considering the faster reduction of Rh oxide in contrast to the slower reoxidation of Rh on ZrP2O7, the perturbation effect can © XXXX American Chemical Society
be rationalized by assuming the stabilization of metallic Rh species, which is more active than Rh oxide for catalytic NO reduction. To demonstrate the hypothesis, in situ experiments for the Rh oxidation state under dynamic TWC conditions at atmospheric pressure are necessary. In situ optical reflectance spectroscopy is known to be useful for detecting the changes in electronic spectra in the UV−vis region under real reaction atmospheres. It is also sensitive for the oxidation state and coordination state of highly dispersed transition metal species.11,12 Weckhuysen et al.12−14 have elucidated the molecular structure of Cr6+ species (via charge transfer transitions) and Cr3+ species (via d−d transitions) on the surface in a CrO3/Al2O3 catalyst during alkane dehydrogenation. Barton et al.15 probed the electronic structure and domain size of different tungsten oxide species on ZrO2. Satsuma et al.16 revealed the redox behavior between W6+ and W(6−δ)+ species supported on ZrO2, which played a key role in alkane skeletal isomerization. They also applied it for the transient kinetic analysis of Rh3+ to Rh0 reduction supported on Received: May 30, 2017 Revised: July 13, 2017 Published: August 4, 2017 A
DOI: 10.1021/acs.jpcc.7b05260 J. Phys. Chem. C XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry C Al2O317 and for the discrimination between Ag+ ions and Ag metal clusters formed in Ag/Al2O3.18,19 Kondratenko et al. studied in situ monitoring of the reduction and reoxidation dynamics of oxygen storage materials based on CeO2−ZrO2.20 In situ optical reflectance spectroscopy has also been used in the analysis of reducibility and/or the oxidation states of V,21,22 Cu,23 Re,24 and Mn.25 However, to the best of our knowledge, the dynamics of Rh redox cycles have not been characterized under TWC reaction conditions. For deeper insight into the redox behavior between Rh3+ and Rh0 and its role in catalysis, the present study applied in situ optical spectroscopy to Rh nanoparticles supported on ZrP2O7 and ZrO2 under simulated TWC gas feeds containing NO, CO, C3H6, and O2 with different A/F values. Because the spectra in the visible light range were changed in accordance with the oxidation states determined by ex situ X-ray photoelectron spectroscopy (XPS), time-resolved experiments were performed for the kinetic analysis of Rh redox behavior. Based on the Rh redox dynamics, the effect of supports (ZrP2O7 and ZrO2) on lean NO activity was discussed. In addition, the redox dynamics were analyzed over perturbations between rich and lean gas feeds with different cycle frequencies, which occur in automotive TWC.
radiation (12 kV). Prior to this measurement, the catalysts were treated with a simulated gas equivalent to A/F = 14.1 and 15.0 containing different concentrations of O2 (0.16 and 0.77% for A/F = 14.1 and 15.0, respectively) and the same concentrations of NO, CO, C3H6, and He (see the Supporting Information, Table S1). The value of A/F was calculated from concentrations of each gas in the gas feed, similar to the method by Tanaka et al.26 First, the gas treatment was performed using A/F = 14.1 at 400 °C for 10 min followed by subsequent rapid cooling in a He gas stream. The samples were transferred to the XPS apparatus using a transfer cell (Thermo Scientific). Before the second XPS experiment, the catalysts were treated in a similar manner but under a gas atmosphere with A/F = 15.0. The C 1s signal at 285.0 eV, which was derived from adventitious carbon, was used as a reference to correct the effects of surface charge. Diffuse reflectance UV−vis spectroscopy experiments were performed using a spectrometer (Jasco V-550) equipped with an in situ gas flow cell comprising a heating stage, a quartz window, a water cooling system, and a gas flow system. Powder catalysts were pressed into pellets (∼7 mm in diameter,