Effect of zirconia support crystal structure on the alloying of rhodium and iridium for the improvement of three-way catalysts

Abstract

Three-way catalysts reduce emissions of air pollutants from automobiles. Rhodium has been the main component of these catalysts. However, because the price of rhodium has recently increased, there is now a desire to save rhodium without sacrificing catalytic performance. Alloying rhodium with other metals is a way to reduce the need for rhodium and improve catalytic performance. However, because rhodium is generally immiscible with other metals, preparation of such catalysts is laborious. In this study, we first demonstrated the effect of the support-crystal-structure of zirconia on the alloying of rhodium and iridium. Monoclinic zirconia with a low specific surface area enabled the formation of rhodium–iridium solid-solution oxides after calcination. These oxides were transformed into metallic random alloys by hydrogen pretreatment. The light-off temperature of the alloy catalyst with approximately one-third the rhodium atoms was similar to that of the pure rhodium catalyst. In situ Fourier transform infrared analysis detected concerted adsorption of carbon monoxide and nitric oxide on alloyed rhodium and iridium, respectively. These adsorbed species reacted smoothly, in contrast to the non-alloyed catalyst. The successful realization of the conventional impregnation-based alloying approach opens new horizons for the design of novel bimetallic catalysts with properties tailored for advanced catalytic applications.

---

1. Introduction

Three-way catalysts can be used to purify exhaust gases from automobiles via simultaneous oxidation and reduction reactions. In recent years, zero-emission vehicles such as battery electric vehicles and fuel-cell electric vehicles have attracted attention. However, it is difficult for these new vehicles to use the existing infrastructural facilities, and their cruising range is still insufficient and dependent on temperature.¹ One report has predicted that the number of these new types of vehicles produced in the future will be only about half that of vehicles with internal combustion engines.² Continued research to improve the performance of the three-way catalysts is therefore desirable from the standpoint of environmental protection. However, there are difficult challenges associated with the continued use of three-way catalysts. The price of rhodium (Rh), which is an essential component of three-way catalysts, recently increased abruptly, and the emission levels permitted in each country will soon be more strict than the current levels.² To overcome these challenges, there is a need to explore novel catalyst technologies that can improve the performance of three-way catalysts with a low Rh content.

Numerous efforts have been made to improve the performance of Rh catalysts. Some researchers, for example, have studied oxide supports with the goal of improving oxygen storage capacity^3,4 and metal–support interactions.^5–9 Other researchers have applied layered Rh catalysts with several additives.^10–13 A recent trend has been alloying Rh with other metals. Alloying causes some of the Rh atoms to be replaced with other metals (i.e., saved), and the effect of alloying (e.g., ensemble and ligand effects) dramatically changes the catalytic performance.¹⁴ For instance, the potential of Rh–Pd nanoalloys, despite their limited miscibility, has been demonstrated to enhance the three-way catalytic performance.¹⁵ However, this type of research is progressing slowly because the immiscibility of Rh with other metals makes it difficult to control the alloy. Previous alloying studies have dealt with this immiscibility problem either by chemical reduction methods using an organic solvent and/or reductant or by impregnation methods using commercially unavailable double-complex salts.^16–19 For industrial applications, a simple and scalable alloying method is highly desirable.

In this study, we first demonstrate the effect of zirconia (ZrO₂) support crystal structure on the alloying of immiscible Rh and iridium (Ir) with simple impregnation method using only commercially available materials. The result was improved three-way catalytic performance with a low Rh content. The phase diagram of a Rh–Ir binary alloy system indicates that equal amounts of Rh and Ir are immiscible below ca. 1300 °C.²⁰ However, a monoclinic ZrO₂ support with a low specific surface area (∼12 m² g⁻¹) enables Rh and Ir to form a solid-solution oxide during calcination because of their similar dispersions in the relatively-low surface area. The formed RhIr solid-solution oxide can be transformed into a metallic random alloy by hydrogen (H₂) pretreatment at temperatures up to 400 °C. The light-off temperatures of the RhIr alloy catalyst were similar to those of a pure Rh catalyst, although the alloy catalyst contained only about one-third of the Rh atoms as the pure Rh catalyst. In situ Fourier transform (FT) infrared (IR) analysis with probe molecules suggested that the improved performance of the RhIr alloy catalyst was induced by the concerted adsorption of carbon monoxide (CO) and nitric oxide (NO) species on the alloy surfaces.

2. Experiment

2.1. Materials

Solutions of Rh and Ir nitrates were purchased from TANAKA Precious Metal (Japan) and ISHIFUKU Metal Industry (Japan), respectively. Yttria-stabilized ZrO₂ (TZ series) was purchased from TOSOH (Japan). The Brunauer–Emmett–Teller (BET) specific surface areas of ZrO₂ containing 0 mol% yttria (0-ZrO₂) and 8 mol% yttria (8-ZrO₂) were 12 and 13 m² g⁻¹, respectively. ZrO₂ with a relatively high specific surface area (100 m² g⁻¹) was purchased from DAIICHI KIGENSO KAGAKU KOGYO (Japan).

2.2. Catalyst preparation

A RhIr alloy catalyst was prepared using a simple impregnation method. A mixed aqueous solution of Rh and Ir nitrates was dropped onto ZrO₂ powder and stirred. The dried powder was calcined at 600 °C for 5 h in flowing air. The total metal loading was 0.5 wt%, and the molar ratio of Rh to Ir was set to 1. The monometallic Rh and Ir catalysts were prepared using the same procedure with a 0.5 wt% metal loading. A physically mixed Rh + Ir catalyst (denoted as Rh + Ir/0-ZrO₂) was prepared by mixing 0.5 wt% Rh/0-ZrO₂ and 0.5 wt% Ir/0-ZrO₂ at a weight ratio of 1.87. This ratio corresponds to the atomic weight ratio of Rh and Ir and was selected to ensure a 1 : 1 molar ratio, which matches that of the RhIr alloy catalyst. This study focused on an equimolar composition because this point tends to be relatively unfavorable thermodynamically for the alloying of immiscible metals. At the same time, it is also the critical point where the local atomic environment is maximally altered by alloying, resulting in a marked change in the structural and electronic properties from those of the original metals.

2.3. Catalyst evaluation

The three-way catalytic performances of the prepared samples were evaluated using a fixed-bed, flow-reactor system. The catalyst powder (50 mg) was loaded into a tubular reactor containing quartz wool. The catalysts were then pretreated with 5% H₂ diluted with nitrogen (N₂) at 400 °C for 30 min, and residual H₂ was then purged with N₂ at the same temperature. After cooling to 50 °C, stoichiometric model gases composed of 0.35% CO, 0.15% NO, 0.33% propylene (C₃H₆), 0.25% oxygen (O₂), and 10% water vapor diluted with N₂ were introduced into the reactor at a total flow rate of 500 mL min⁻¹ (SV = 600 000 h⁻¹). The catalysts were then heated in increments of 10 °C min⁻¹ from 50 to 600 °C, and the effluent gases were continuously monitored using two gas analyzers (PG-340, Horiba, Japan and VMS-1000F, Shimadzu, Japan). Note that the catalyst weights used in the catalyst evaluation of monometallic and bimetallic catalysts were the same, but the metal contents were different (Table S1†).

3. Results and discussion

3.1. Dispersibility of Rh and Ir on monoclinic and cubic ZrO₂

Fig. 1a illustrates the process of catalyst preparation. The impregnated Rh and Ir species were crystallized via calcination and H₂ pretreatment before evaluation of the catalyst. Because the dispersibility of the crystallized Rh and Ir particles directly affected the performance of the three-way catalyst, we first compared the X-ray diffraction (XRD) patterns of the calcined catalysts. The 0-ZrO₂ support exhibited a pure diffraction pattern attributable to its monoclinic structure (Fig. 1b) that did not change significantly after metal loading. The indication was that the amount of loaded metals was insufficient for detection and/or that the supported metals were highly dispersed on the monoclinic 0-ZrO₂. In contrast, the diffraction pattern of the 8-ZrO₂ support was attributable to its pure cubic structure, and Ir/8-ZrO₂ and RhIr/8-ZrO₂ showed a small peak at approximately 2θ = 28° (Fig. 1c). The implication was that the supported Ir species had crystallized to form aggregate IrO₂ particles on the cubic 8-ZrO₂.

Fig. 1 (a) Schematic of the preparation of the catalyst. XRD patterns of calcined catalysts using (b) 0-ZrO₂ and (c) 8-ZrO₂ supports. (d) High-angle annular dark-field (HAADF) image and EDXS maps of calcined RhIr/monoclinic 0-ZrO₂.

To quantify the dispersibility of the supported metals, CO chemisorption experiments were performed for the monometallic catalysts (Fig. S1†). After H₂ pretreatment, monometallic Rh/monoclinic 0-ZrO₂ and Ir/monoclinic 0-ZrO₂ exhibited comparable dispersion. In contrast, Rh and Ir were highly and poorly dispersed, respectively, on cubic 8-ZrO₂. These results indicated that the dispersibilities of Rh and Ir depended strongly on the crystal structure of the ZrO₂ support and were consistent with the XRD results. To explore the origin of the support-structure-dependent metal dispersibility, carbon dioxide (CO₂) temperature-programmed desorption (CO₂-TPD) measurements were performed (Fig. S2†). Previous studies have reported that yttria doping to ZrO₂ increases the number of basic sites via the formation of oxygen vacancies.²¹ In our study, the non-supported 0-ZrO₂ and 8-ZrO₂ exhibited desorption peaks at approximately 150 °C and 300 °C. The latter peak showed slightly increased intensity in 8-ZrO₂, which may be attributed to basic sites formed through oxygen vacancies induced by yttria doping. However, the overall difference in surface basicity between the two supports was insufficient to explain the pronounced contrast in metal dispersibility observed in the CO chemisorption experiments. In contrast, high-resolution scanning transmission electron microscopy (STEM) imaging revealed that small metal nanoparticles on 8-ZrO₂ had exhibited an epitaxial-like alignment with the cubic ZrO₂ lattice (Fig. S3†). The measured lattice spacing of approximately 0.22 nm corresponded to the (111) plane of Rh; however, this value was also close to that of Ir. Nevertheless, given that Ir had shown a strong tendency to aggregate on cubic 8-ZrO₂ as confirmed by XRD and CO chemisorption experiments, these epitaxially aligned nanoparticles were most likely composed of Rh. This finding suggested that the high dispersion of Rh on 8-ZrO₂ had primarily resulted from favorable lattice matching between Rh and the cubic ZrO₂. No such alignment was observed for Ir, indicating that the support-structure dependence of metal dispersibility was likely governed more by interfacial crystallographic compatibility than by differences in surface basicity. In the case of bimetallization, this effect of crystal structure on the dispersibility of Rh and Ir may also have affected the atomically close-contact state of these metals on the low-specific-surface-area support (∼13 m² g⁻¹) used in this study.

We then used STEM combined with energy-dispersive X-ray spectroscopy (EDXS) to visualize the atomically close-contact states of the supported Rh and Ir species in the bimetallic RhIr/monoclinic 0-ZrO₂ and cubic 8-ZrO₂. The RhIr/monoclinic 0-ZrO₂ exhibited a supported particle with dimensions of approximately 10 nm (Fig. 1d). From this particle, Rh-L and Ir-L signals, which corresponded to approximately equivalent moles, were detected in the EDXS analysis (Fig. S4†). The implication was that the Rh and Ir species were atomically mixed and close together in a single particle. In contrast, RhIr/cubic 8-ZrO₂ exhibited small Rh particles and large Ir particles that were independent of one another (Fig. S5†). This independence was consistent with the XRD and CO chemisorption results. These results indicated that the crystal structure of ZrO₂ strongly affected the dispersibility of the supported Rh and Ir species in both monometallic and bimetallic systems, and it changed the close-contact state of the supported Rh and Ir species in the bimetallic systems.

3.2. Structural characterization of RhIr composite formed on monoclinic 0-ZrO₂

We used X-ray absorption spectroscopy to investigate the structure of the RhIr composite formed on monoclinic 0-ZrO₂. Fig. 2a and b show the k³-weighted Rh K- and Ir L₂-edge FT extended X-ray absorption fine structure (EXAFS) spectra of calcined RhIr/monoclinic 0-ZrO₂ and RhIr/cubic 8-ZrO₂. Fig. S6† shows the X-ray absorption near-edge structure (XANES) spectra and EXAFS oscillations before FT. In both edges in Fig. 2, there were peaks at around 1.5 Å in the spectra of RhIr/monoclinic 0-ZrO₂ and RhIr/cubic 8-ZrO₂ attributable to Rh–O or Ir–O scattering.^22,23 The implication was that the Rh and Ir species supported on monoclinic 0-ZrO₂ and cubic 8-ZrO₂ were oxides after calcination. The second-shell peaks at ∼2.5 Å in the Rh K-edge disappeared only in RhIr/monoclinic 0-ZrO₂. The indication was that the Rh species in the RhIr composite that formed on monoclinic 0-ZrO₂ did not have the original (i.e., Rh₂O₃) structure, in contrast to RhIr/cubic 8-ZrO₂. In contrast, the Ir L₂-edge FT EXAFS of the RhIr/monoclinic 0-ZrO₂ was similar to that of the RhIr/cubic 8-ZrO₂ and IrO₂ reference, but the second-shell peaks at ∼3 Å were shifted to shorter distances. Based on these results, we proposed that the RhIr composite formed on monoclinic 0-ZrO₂ was likely a solid-solution oxide, possibly involving the substitution of Rh into a rutile-type IrO₂ lattice. This interpretation was further supported by the XANES results: the Rh K-edge of RhIr/monoclinic 0-ZrO₂ exhibited a noticeable shift toward higher energy than that of RhIr/cubic 8-ZrO₂ and Rh₂O₃, suggesting a more oxidized Rh state (Fig. S6a†). Similarly, the Ir L₂-edge showed an increase in the white line intensity relative to RhIr/cubic 8-ZrO₂ and IrO₂, indicating a change in the electronic environment around Ir (Fig. S6b†). These features were consistent with the mutual electronic interactions between Rh and Ir species, suggesting the possible formation of a solid solution oxide. While the specific substitution of Rh into the rutile-type IrO₂ lattice remains a working hypothesis, overall spectroscopic evidence supports the notion that Rh and Ir form a solid-solution structure on monoclinic ZrO₂.

Fig. 2 (a) Rh K-edge and (b) Ir L₂-edge FT EXAFS spectra of calcined RhIr/monoclinic 0-ZrO₂ and RhIr/cubic 8-ZrO₂.

3.3. Process of transforming RhIr solid-solution oxide to metallic random alloy on monoclinic 0-ZrO₂

The prepared and calcined catalysts were pretreated with H₂ before evaluation of the catalysts (Fig. 1a). We then measured the H₂-temperature-programmed reduction (TPR) profile of each catalyst after calcination (Fig. 3). Monometallic Rh and Ir/monoclinic 0-ZrO₂ exhibited peaks derived from H₂ consumption at approximately 100 °C and 200 °C, respectively. A physical mixture of these monometallic catalysts also exhibited peaks at approximately 100 °C and 200 °C. These results indicated that stepwise consumption of H₂ occurred in the presence of pure Rh and Ir oxides. Stepwise consumption of H₂ was also observed in the case of RhIr/cubic 8-ZrO₂, where solely supported Rh and Ir oxides were detected by STEM–EDXS and FT-EXAFS analyses. The change of the peak position between the physical mixture of the monometallic catalysts and RhIr/cubic 8-ZrO₂ may have been due to the difference in the dispersibility of Rh and Ir on monoclinic and cubic ZrO₂. In contrast, RhIr/monoclinic 0-ZrO₂ exhibited a single peak at ∼150 °C. The implication was that Rh and Ir species in the solid-solution oxide were simultaneously reduced to metals. A previous alloying study with chemical reduction has reported that simultaneous co-reduction of metal cations is a key step in the production of metallic random alloys.²⁴ It is thus possible that RhIr solid-solution oxide was transformed into a metallic random alloy via simultaneous reduction during H₂-pretreatment in this study.

Fig. 3 H₂-TPR profiles of monometallic and bimetallic catalysts. Before measurement, the catalysts were pretreated with 5% O₂ diluted with argon (Ar) at 600 °C for 10 min, and then the residual O₂ was purged with Ar at the same temperature.

To experimentally detect the formation of a metallic random alloy, we collected H₂-pretreated RhIr/monoclinic 0-ZrO₂ and analyzed it using FT-EXAFS and STEM–EDXS. Fig. 4a and b show the k³-weighted Rh K- and Ir L₂-edge FT-EXAFS spectra of H₂-pretreated RhIr/monoclinic 0-ZrO₂. Fig. S7† shows the XANES spectra and EXAFS oscillations before FT. In both edges in Fig. 4a and b, the FT EXAFS spectra of the H₂-pretreated RhIr/monoclinic ZrO₂ differed from those of the metal and oxide references. The split peaks at 2–3 Å in the Rh K- and Ir L₂-edge were successfully fitted using the Rh–Rh, Rh–O, Rh–Ir, Ir–Ir, Ir–O, and Ir–Rh scattering paths (Fig. S8†). The indication was that the observed split peaks originated from the formation of RhIr alloys. STEM–EDXS maps of H₂-pretreated RhIr/monoclinic 0-ZrO₂ showed that the Rh-L and Ir-L signals overlapped and were randomly distributed in a single particle. The indication was that the RhIr alloys supported on monoclinic 0-ZrO₂ had the structure of a random alloy (Fig. 4c). The Rh–Ir system is classically known to be immiscible across a wide composition range under equilibrium conditions, as reported by Raub.²⁵ Our experimental evidence is the first demonstration that generally immiscible Rh and Ir can be alloyed by facile impregnation methods that exploit the effect of the crystal structure of ZrO₂. On monoclinic ZrO₂ with a low specific surface area (12 m² g⁻¹), supported Rh and Ir form solid-solution oxides because they are similarly dispersed in its relatively low surface area. These incorporated Rh and Ir species can be simultaneously reduced by H₂-pretreatment at ∼150 °C to form RhIr metallic random alloys. For comparison, RhIr/monoclinic ZrO₂ was prepared using a ZrO₂ support with a relatively high specific surface area (100 m² g⁻¹). However, the H₂-TPR profile of this catalyst exhibited two peaks derived from stepwise consumption of H₂ (Fig. S9†). The implication was that the monoclinic ZrO₂ support with its high specific surface area did not facilitate the alloying mechanism proposed in this study. Compared to the H₂-TPR profile of RhIr/0-ZrO₂ with a low specific surface area, the H₂ consumption peaks shifted to higher temperatures and became broader. These features suggested the presence of heterogeneous and stronger metal–support interactions, likely arising from the increased surface defects and/or hydroxyl groups associated with a higher surface area. Such enhanced interactions may interfere with Rh–Ir alloy formation by relatively weakening the interaction between Rh and Ir, thereby hindering the alloying.

Fig. 4 (a) Rh K-edge and (b) Ir L₂-edge FT EXAFS spectra of H₂-pretreated RhIr/monoclinic 0-ZrO₂. (c) HAADF image and EDXS maps of H₂-pretreated RhIr/monoclinic 0-ZrO₂.

3.4. Three-way catalytic performance of RhIr alloys supported on monoclinic 0-ZrO₂

Finally, we investigated the effect of alloying Rh and Ir on three-way catalytic performance. All the catalysts were pretreated with 5% H₂ diluted with N₂ at 400 °C for 30 min and then tested under stoichiometric gas conditions. To eliminate the influence of transient surface states after H₂ pretreatment, the light-off performance was evaluated based on the second heating cycle. For the representative alloyed RhIr catalyst, repeated measurements confirmed that the catalytic activity was stabilized from the second cycle (Fig. S10†). In the monometallic system, Rh/monoclinic 0-ZrO₂ exhibited a better light-off curve for all gases than Ir/monoclinic 0-ZrO₂ (Fig. S11†). Because the metal dispersions of these monometallic catalysts were comparable (Fig. S1†), Rh appears to be inherently active in the three-way catalytic reaction. However, because the atomic amount of Rh is approximately half that of Ir, twice as much Rh was used in the evaluation of the Rh catalysts (Table S1†). To eliminate the influence of such differences in metal content on catalytic performance, we compared alloyed RhIr/monoclinic 0-ZrO₂ and physically mixed Rh + Ir/monoclinic 0-ZrO₂. Interestingly, the light-off temperatures of the alloy catalysts were lower than those of the physically mixed catalysts for all the gases (Fig. 5). The metal dispersion of alloyed RhIr/monoclinic 0-ZrO₂ was also comparable to that of the physically mixed Rh + Ir/monoclinic 0-ZrO₂ (Fig. S12†). The indication was that the effect of alloying was to accelerate the three-way catalytic reaction. Notably, the light-off temperatures of the alloyed RhIr/monoclinic 0-ZrO₂ were lower for C₃H₆ conversion than that of monometallic Rh/monoclinic 0-ZrO₂, although the alloy catalyst contained approximately one-third as many Rh atoms as the monometallic Rh catalyst (Fig. S13†). In other words, the beneficial alloying effect enabled us to reduce the amount of Rh to 1/3 with minimal sacrifice in the catalytic performance. Moreover, STEM–EDXS analysis of the RhIr/monoclinic 0-ZrO₂ catalyst after the second light-off test showed an overlapping of Rh-L and Ir-L signals within individual particles, similar to the fresh catalyst although the redox state of the catalyst after the reaction could not be definitively determined (Fig. S14†). This observation implied that the solid-solution structures of Rh and Ir were retained even after the reaction.

Fig. 5 Three-way catalytic performance of the alloyed RhIr/monoclinic 0-ZrO₂ and physically mixed Rh + Ir/monoclinic 0-ZrO₂. Before the evaluation, the catalysts were pretreated with 5% H₂ diluted with N₂ at 400 °C for 30 min, and then the residual H₂ was purged with N₂ at the same temperature.

To identify the effect of alloying on the three-way catalytic performance, we performed in situ FT-IR analysis using CO and NO as probe molecules at 150 °C. In this experiment, NO was introduced to CO-saturated catalyst surfaces because CO was adsorbed faster than NO when H₂-pretreated and alloyed RhIr/monoclinic 0-ZrO₂ was exposed to mixed reaction gases (Fig. S15†). Preliminary experiments with monometallic Rh and Ir/monoclinic 0-ZrO₂ were performed to facilitate peak assignments. In the monometallic Rh system, peaks assigned to Rh(CO), Rh(CO)₂, and Rh(NO) appeared at 2063, 2085/2018, and 1838 cm⁻¹, respectively (Fig. S16a†).²⁶ In the monometallic Ir system, the peaks assigned to Ir(CO) and Ir(NO) appeared at 2076 and 1795 cm⁻¹, respectively (Fig. S16b†).²⁷ Naturally, all the adsorbed species detected in this monometallic system were observed on the physically mixed Rh + Ir/monoclinic 0-ZrO₂, although it was difficult to distinguish the Rh(CO) and Ir(CO) peaks because they overlapped (Fig. 6a). However, the Rh(NO) peak disappeared in the IR spectra of the alloyed RhIr/monoclinic 0-ZrO₂, and the remaining Ir(NO) peak gradually decreased after the introduction of NO was stopped and NCO species began to be formed via the CO + NO reaction (Fig. 6b).²⁶ In contrast, the NO adsorbed on the Ir sites of the physically mixed catalysts was relatively inert and remained even after the introduction of NO was stopped (Fig. 6a). These results indicated that NO was preferentially adsorbed on the Ir sites on the surface of the alloy, whereas Rh was adsorbed only on CO. This concerted adsorption may accelerate the CO + NO reaction over the surface of the alloy. Although this concerted mechanism may facilitate the ignition of the reaction at lower temperatures, it also requires surface diffusion and the subsequent reaction of adsorbed intermediates between spatially separated Rh and Ir sites. These spatial and kinetic constraints may limit the acceleration of the reaction rate at higher temperatures, leading to a more gradual increase in conversion, which was observed as a gentler slope in the light-off curves of the alloyed catalyst compared to the physically mixed catalyst (Fig. 5). Based on these findings, we concluded that the enhancement of the three-way catalytic performance through alloying of Rh and Ir originated primarily from the preferential adsorption of NO on Ir sites and the subsequent increase of the reactivity of the alloy. Accelerating the NO conversion also improved the conversions of CO and C₃H₆ as reductants.

Fig. 6 IR spectra recorded at 150 °C after sequential flowing of mixed gases composed of 0.15% CO + 0.15% NO and 0.15% CO diluted with Ar through (a) physically mixed Rh + Ir/monoclinic 0-ZrO₂ and (b) alloyed RhIr/monoclinic 0-ZrO₂. Before the measurement, the catalysts were pretreated with 5% H₂ diluted with Ar at 400 °C for 30 min, and then the residual H₂ was purged with Ar at the same temperature.

4. Conclusions

The goal of this study was to improve the performance of a three-way catalyst with low Rh content by alloying Rh and Ir. Although Rh and Ir are phase-separated during general catalyst preparation, this immiscibility problem was overcome for the first time by the effect of the support-crystal-structure of ZrO₂. Monoclinic ZrO₂ with a low specific surface area enabled Rh and Ir to be supported with a similar dispersion. The result was formation of RhIr solid-solution oxides after calcination. Monoclinic ZrO₂-supported RhIr solid-solution oxides were successfully transformed into metallic random alloys by H₂ pretreatment up to 400 °C. The proposed alloying procedure was based completely on the facile impregnation method. The light-off temperatures of the RhIr alloy catalyst with approximately one-third the number of Rh atoms were similar to those of the pure Rh catalyst. In situ FT-IR analysis revealed that atomically alloyed Rh and Ir preferentially adsorbed CO and NO, respectively, and this concerted adsorption contributed to the improved performance of the three-way catalyst.

Data availability

The data supporting this article have been included as part of the ESI.†

Author contributions

Yoshihide Nishida: designing of research, data collection, and writing – editing. Koki Aono: conducting STEM–EDXS analysis. Hirona Yamagishi, Hiromi Togashi, and Shunsuke Oishi: conducting X-ray absorption spectroscopy. Masaaki Haneda: designing of research.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

We acknowledge writing support from Kohei Kusada and Hiroshi Kitagawa of Kyoto University. X-ray absorption analyses were carried out at beamline BL01B1 in SPring-8 (proposal no. 2023B1591) and at BL11S2 in Aichi SR (proposal no. 202203074, 202204060, 202206047).

References

1. Y. Omori, A. Shigemoto, K. Sugihara, T. Higo, T. Uenishi and Y. Sekine, Catal. Sci. Technol., 2021, 11, 4008–4011.
2. S. P. Denisov, E. A. Alikin, E. O. Baksheev and V. N. Rychkov, Catal. Ind., 2023, 15, 443–449.
3. J. Wu, A. E. O'Neill, C.-H. Li, J. R. Jinschek and G. Cavataio, Appl. Catal., B, 2021, 280, 119450.
4. K. Fujita, H. Asakura, S. Hosokawa, K. Teramura, M. Kobayashi, K. Fujita and T. Tanaka, ACS Appl. Mater. Interfaces, 2021, 13, 24615–24623.
5. C.-H. Li, J. Wu, A. B. Getsoian, G. Cavataio and J. R. Jinschek, Chem. Mater., 2022, 34, 2123–2132.
6. M. Machida, S. Iwashita, T. Sato, H. Yoshida, J. Ohyama, M. Yoshioka, S. Miwa and T. Hashishin, J. Phys. Chem. C, 2022, 126, 17608–17617.
7. M. Machida, H. Yoshida, N. Kamiuchi, Y. Fujino, T. Miki, M. Haneda, Y. Tsurunari, S. Iwashita, R. Ohta, H. Yoshida, J. Ohyama and M. Tsushida, ACS Catal., 2023, 13, 3806–3814.
8. L. Zeng, J. Fan, J. Mou, L. Chen, Z. Song, P. Liu, Y. Jiao, J. Wang and Y. Chen, J. Energy Inst., 2023, 111, 101436.
9. L. Chen, J. Fan, P. Liu, J. Mou, L. Zeng, Z. Song, Y. Jiao, J. Wang and Y. Chen, Appl. Catal., A, 2023, 668, 119494.
10. H. Yoshida, Y. Kuzuhara, T. Koide, J. Ohyama and M. Machida, Catal. Today, 2021, 375, 258–263.
11. H. Yoshida, T. Koide, T. Uemura, Y. Kuzuhara, J. Ohyama and M. Machida, Catal. Today, 2023, 410, 102–108.
12. M. Machida and H. Yoshida, Mater. Trans., 2023, 64, 2369–2375.
13. H. Yoshida and M. Machida, ChemCatChem, 2023, 15, e202300771.
14. H. Miura and T. Shishido, Chem. Lett., 2021, 50, 346–352.
15. A. A. Vedyagin, Y. V. Shubin, R. M. Kenzhin, P. E. Plyusnin, V. O. Stoyanovskii and A. M. Volodin, Top. Catal., 2019, 62, 305–314.
16. H. Guo, Z. Fang, H. Li, D. Fernandez, G. Henkelman, S. M. Humphrey and G. Yu, ACS Nano, 2019, 13, 13225–13234.
17. H. Minamihara, K. Kusada, T. Yamamoto, T. Toriyama, Y. Murakami, S. Matsumura, L. S. R. Kumara, O. Sakata, S. Kawaguchi, Y. Kubota, O. Seo, S. Yasuno and H. Kitagawa, J. Am. Chem. Soc., 2023, 145, 17136–17142.
18. X. Zhou, M. Mukoyoshi, K. Kusada, T. Yamamoto, T. Toriyama, Y. Murakami, S. Kawaguchi, Y. Kubota, O. Seo, O. Sakata, T. Ina and H. Kitagawa, Chem. Sci., 2024, 15, 7560–7567.
19. A. A. Vedyagin, V. O. Stoyanovskii, R. M. Kenzhin, P. E. Plyusnin and Y. V. Shubin, Dokl. Phys. Chem., 2022, 506, 131–137.
20. S. N. Tripathi, S. R. Bharadwaj and M. S. Chandrasekharaiah, J. Phase Equilib., 1991, 12, 606–608.
21. A. S. Al-Fatesh, A. A. Ibrahim, A. H. Fakeeha, F. Albaqi, K. Anojaidi, I. Albinali, A. E. Abasaeed, F. Frusteri, S. L. Mahmud, J. K. Abu-Dahrieh and A. A. Bagabas, ACS Omega, 2022, 7, 43700–43709.
22. H. Asakura, S. Hosokawa, T. Ina, K. Kato, K. Nitta, K. Uera, T. Uruga, H. Miura, T. Shishido, J. Ohyama, A. Satsuma, K. Sato, A. Yamamoto, S. Hinokuma, H. Yoshida, M. Machida, S. Yamazoe, T. Tsukuda, K. Teramura and T. Tanaka, J. Am. Chem. Soc., 2018, 140, 176–184.
23. J. Fu, J. Dong, R. Si, K. Sun, J. Zhang, M. Li, N. Yu, B. Zhang, M. G. Humphrey, Q. Fu and J. Huang, ACS Catal., 2021, 11, 1952–1961.
24. K. Kusada, D. Wu and H. Kitagawa, Chemistry, 2020, 26, 5105–5130.
25. E. Raub, J. Less-Common Met., 1959, 1, 3–18.
26. J. Floresmoreno, G. Delahay, F. Figueras and B. Coq, J. Catal., 2005, 236, 292–303.
27. Y. Bai, S. Gao, Y. Sun, W. Ouyang, Y. Zhou, H. Wang and Z. Wu, Environ. Sci. Technol., 2023, 57, 9105–9114.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5cy00561b

‡ Masaaki Haneda died on May 26, 2024.

---