Synthesis of supported immiscible nanoalloy catalysts via gas-switching reduction in the impregnation method

Abstract

Supported catalysts are widely used in chemical processes. For preparing supported catalysts, impregnation methods are recognized as straightforward procedures. Although the development of high-performance supported catalysts using impregnation methods is under consideration for various chemical reactions, there is a need for technologies that improve catalytic performance by exploiting the divergent physicochemical properties of active metals. In this study, we focused on alloying active metals to exploit their divergent properties and produced a simple impregnation-based alloy of immiscible rhodium, palladium, and platinum supported on alumina. Previous alloying methods have depended on elaborate and rigid procedures to control crystallization. However, we integrated the key alloying principles found in previous studies into a heat-treatment process of impregnation and developed a “gas-switch-triggered reduction method” in which all metal cations are simultaneously reduced by simply switching the treatment gas at a certain temperature. We applied the developed method to the alumina-supported ternary rhodium–palladium–platinum system. Alloying active metals led to 18 times higher catalytic performance in nitrile hydrogenation than that of the monometallic catalysts. The developed method requires no special equipment or procedures such as those used in previous studies and can be merged into the pretreatment process before catalyst evaluation using a continuous gas flow system, where in situ alloy formation occurs without the oxidation of the alloyed metals because of the absence of air exposure. The successful realization of the simple impregnation-based alloying method accelerates catalyst design based on the differences of the crystalline nature by random alloying and provides a bridge to their industrial application.

---

Introduction

Supported catalysts are widely used in various chemical processes. Several methods are available for preparing supported catalysts, including the impregnation method, which is a straightforward procedure in which metal precursors and oxide supports are mixed, dried, and crystallized by heat treatment under certain gas conditions.¹ Although this method has produced high-performance catalysts for various chemical reactions,^2–5 a catalyst that performs well in one reaction rarely performs well in another reaction. Catalysts must therefore be customized for each chemical reaction. However, the available metal and oxide support species for supported catalysts are limited in the periodic table, and the simplicity of the impregnation method restricts the variety of supported catalysts that can be synthesized. To meet the expanding demand for catalytic reactions driven by future societal and industrial transformations, it is essential to explore technologies that diversify and improve the performance of supported catalysts.

Random alloying involves the mixing of atoms of metal A with atoms of metal B. The result is a dramatic change of the original crystalline nature of the metals. A notable example is the formation of pseudo-rhodium (Rh) by the alloying of neighboring ruthenium (Ru) and palladium (Pd), as reported by Kusada et al.^6,7 However, the alloying strategy has often encountered challenges owing to metal immiscibility. Previous alloying studies have dealt with this immiscibility problem using elaborate and/or low-scale methods to control the crystallization process.^8–11 For industrial applications, alloying by simple impregnation using commercially available materials and without special equipment and/or procedures is desirable. A method closer to this target is the alloying of immiscible metals using hydrogen spillover on a reducible oxide support in a simple impregnation method, as reported by Mori et al.^12–14

In this study, we first demonstrated the simple impregnation-based alloying of an immiscible ternary Rh–Pd–Pt (ref. 15) system supported on non-reducible alumina (Al₂O₃). Alloying was successfully achieved by integrating the core principle of the simultaneous co-reduction of all metal cations, known from previous alloying studies,^8,9 into the heat-treatment step of a simple impregnation method. The treatment gas was simply switched from inert argon (Ar) to hydrogen (H₂) at the reducible temperature of all supported metal cations, and it was only necessary that there be switching valves in the inert gas and H₂ gas lines, which are likely to be present. This gas-switch-triggered reduction method can also be merged into the pretreatment process before catalyst evaluation using a continuous gas flow system, where in situ alloy formation occurs without the oxidation of the alloyed metals because of the absence of air exposure. In the hydrogenation of nitriles that we tested as a model reaction, the ternary alloyed Rh–Pd–Pt catalysts exhibited catalytic performance up to 18 times that of the monometallic catalysts.

Experimental

Material

Rh, Pd, and Pt precursors were purchased from Sigma-Aldrich (USA) and FUJIFILM Wako Pure Chemical Corporation (Japan). High-purity γ-Al₂O₃ (AKP-G15) and silica (SiO₂) were supplied by SUMITOMO CHEMICAL COMPANY (Japan) and FUJI SILYSIA CHEMICAL (Japan). The Brunauer–Emmett–Teller (BET) specific surface area of γ-Al₂O₃ and SiO₂ were 164 and 300 m² g⁻¹, respectively. For evaluation of the catalyst, benzonitrile as a substrate, dodecane as an internal standard, and methanol (MeOH) as a solvent were purchased from Tokyo Chemical Industry (Japan).

Supported nanoalloy synthesis

All the catalysts were prepared using a simple impregnation method. Aqueous solutions of RhCl₃, (NH₄)₂PdCl₄, and/or (NH₄)₂PtCl₄ were dropped onto γ-Al₂O₃ powder, stirred, and dried. One type of metal precursor was used for the monometallic system and three types for the trimetallic system. Unless otherwise noted, all multiple-metallic catalysts were synthesized with an equimolar ratio of the respective metal precursors. The total metal loading was set to 3 weight percent (wt%). In the heat treatment process, the dried powder was loaded into a tubular quartz reactor with quartz wool. Inert Ar was introduced into the catalyst bed, and the catalyst was heated in increments of 10 °C min⁻¹ from room temperature to 600 °C. After reaching 600 °C, the treatment gas was switched from Ar to H₂, and the temperature was maintained for 180 min. Finally, the residual H₂ was purged with Ar at 600 °C for 30 min, and the resulting powder was cooled to room temperature under flowing Ar and exposed to atmospheric air to collect it in vials. To further investigate the generality of the method, we additionally synthesized PdPt/Al₂O₃, RhPdPt/SiO₂, and RhPdPt/Al₂O₃ with varied compositions (6 : 2 : 2, 2 : 6 : 2, and 2 : 2 : 6, molar ratios). These catalysts were prepared using the same impregnation and heat treatment procedure described above. For comparison, the catalyst was heat-treated while supplying H₂ from the initial stage of heating, as in the conventional impregnation method.¹⁶

Catalyst evaluation

Nitrile hydrogenation was performed in a test tube equipped with a gas bag as shown in our previous work.¹⁷ In brief, the tested catalysts were not pretreated after synthesis. In a typical reaction, 1 mmol of benzonitrile and 1 mol% catalyst was dispersed in 1.0 mL of MeOH. The test tube was then depressurized and filled with H₂ gas at atmospheric pressure. The solution was stirred at 1200 rpm using a magnetic stirrer, and the reaction was allowed to proceed at 25 °C. Nitrile conversion and product yield were calculated by gas chromatography with flame ionization detection using dodecane as an internal standard.

Results and discussion

Effect of gas switching on alloying of Rh–Pd–Pt supported on γ-Al₂O₃

Fig. 1 illustrates the concept of alloying via gas-switch-triggered reduction in a simple impregnation method. The core principle of alloying immiscible metals has been reported to involve the simultaneous co-reduction of all metal cations,^8,9 which must be integrated into the impregnation method to streamline the alloying process. In contrast, conventional heat treatment in the impregnation method involves increasing the temperature while initially supplying H₂ gas to the dried catalyst powder (Fig. 1, top).¹⁶ The result is the sequential reduction of easily reducible metals in multiple-metallic systems that hinders random alloy formation. To circumvent this sequential reduction, we have developed a method in which an inert gas (e.g., Ar) is supplied to the catalyst powder during the increase of temperature, and the gas supply is switched to H₂ only at a temperature sufficiently high (e.g., 600 °C) that all metal cations can be reduced simultaneously (Fig. 1, bottom). This type of gas switching can be easily performed by introducing a 2-position, 4-port valve into the H₂ and inert gas lines. The images at the top and bottom right side of Fig. 1 show overlays of the scanning transmission electron microscopy (STEM) combined with energy-dispersive X-ray spectroscopy (EDXS) maps of the γ-Al₂O₃-supported Rh–Pd–Pt catalysts treated by conventional heating and the gas-switch-triggered reduction method, respectively. Full images, including the high-angle annular dark-field (HAADF) image, and particle-size distributions are shown in Fig. S1 and S2. The observation of isolated Rh, Pd, and Pt particles in the case of conventional heating (Fig. 1, top) indicated that reduction by conventional heating ineffectively alloyed these metals. In contrast, the clear overlapping of the Rh-L, Pd-L, and Pt-L signals in the image of the catalyst prepared by the gas-switch-triggered reduction method (Fig. 1, bottom) indicated that these metals were mixed in a single particle. In addition, X-ray absorption spectroscopy (XAS), described later in this paper (Fig. 3), suggested that the metals supported on γ-Al₂O₃ prepared by the gas-switch-triggered reduction method did not have their original structures and were alloyed, whereas those in the conventionally reduced sample largely retained their individual characteristics, indicating insufficient alloying. Small and isolated Rh particles were apparent in the bottom EDXS map in Fig. 1. These particles may have originated from the oxidation of the alloyed metals due to atmospheric exposure after synthesis of the alloy, as described later in this paper. These experimental results are the first demonstration that switching the treatment gas enables immiscible Rh–Pd–Pt (ref. 15) to alloy in a simple impregnation method, wherein the metal precursors and oxide support are mixed, dried, and heat-treated without any special equipment or procedures. To further validate the general applicability of the gas-switch-triggered reduction method, additional catalysts were prepared, including bimetallic PdPt/Al₂O₃ and trimetallic RhPdPt/SiO₂, and RhPdPt/Al₂O₃ with varying metal compositions. STEM-EDXS and XAS analyses revealed that PdPt/Al₂O₃ exhibited clear alloy formation (Fig. S3–S6), even though Pd and Pt are generally immiscible below 770 °C according to the bulk phase diagram.¹⁸ In contrast, for RhPdPt/SiO₂, although Pd and Pt were alloyed, Rh remained largely isolated (Fig. S7–S10). This incomplete alloying of Rh may be attributed to differences in the dispersion or anchoring state of the metal precursors, which could result from variations in the physical structure and/or surface properties of the SiO₂ support. On the other hand, RhPdPt/Al₂O₃ with different compositions exhibited alloy formation, as evidenced by the co-location of elements in a single particle (Fig. S11–S13). However, elemental segregation was consistently observed across these samples, likely reflecting the influence of the initial compositional imbalances. Note that XAS analysis was not performed for the composition-varied RhPdPt/Al₂O₃ due to insufficient edge jumps at the Rh and Pd K-edges. Because of the relatively low loading of each individual metal in these trimetallic catalysts, increasing the sample amount to enhance the signal intensity led to excessive background absorption, compromising the spectral quality. Therefore, only STEM-EDXS analysis was used to evaluate the alloying state of these samples. These observations indicate the general applicability of the gas-switch-triggered reduction method across different compositions and systems, while highlighting potential limitations under certain support or compositional conditions. However, all the catalysts in this study were synthesized under the same preparation conditions. Further optimization of parameters such as precursor type, loading amount, reduction temperature, and duration may overcome these limitations and expand the scope of alloy formation.

Fig. 1 Schematic of alloying metals supported on non-reducible γ-Al₂O₃via gas switching in a simple impregnation method. The inserted images show the STEM-EDXS maps of RhPdPt/Al₂O₃ treated by (top) the conventional heating method and (bottom) the gas-switch-triggered reduction method.

Chemical state and stability of Rh, Pd, and Pt precursors supported on γ-Al₂O₃

It has been reported that the use of multiple complex salts as metal precursors, which are commercially unavailable, facilitates alloy formation in the impregnation method.¹⁹ However, multiple complex salts can be formed by pH-controlled sequential adsorption of metal precursors on oxide supports, which results in alloy formation.²⁰ To identify whether the alloying mechanism in the gas-switch-triggered reduction method involved the formation of multiple complex salts, we used XAS analysis to investigate the changes in the chemical states of the γ-Al₂O₃-supported Rh, Pd, and Pt precursors. In this study, we used RhCl₃, (NH₄)₂PdCl₄, and (NH₄)₂PtCl₄ as metal precursors. The distinct coordination environment of RhCl₃ compared to the ammonium salts of Pd and Pt would be expected to cause significant changes in the Rh K-edge X-ray absorption near-edge structure (XANES) spectrum if multiple complex salts were formed. However, the XANES spectra of untreated RhPdPt/Al₂O₃ closely resembled those of the untreated monometallic Rh, Pd, or Pt/Al₂O₃ catalysts (Fig. S14). The absence of edge shifts or changes in the white line intensities suggested that the Rh, Pd, and Pt precursors in RhPdPt/Al₂O₃ remained as isolated species on γ-Al₂O₃ without forming multiple complex salts.

In addition, we investigated the thermal stabilities of the Rh, Pd, and Pt precursors in RhPdPt/Al₂O₃ using thermogravimetric-differential thermal analysis (TG-DTA). The 3 wt% total metal loading in RhPdPt/Al₂O₃ made it challenging to assess the thermal stability of the metal precursors using the TG profiles (Fig. 2 left), but the DTA profiles effectively captured the thermal stability during the heating process because of the characteristic endothermic or exothermic events linked to the thermal decomposition or phase transitions of the metal precursors. For example, under H₂ flow, monometallic Rh/Al₂O₃ exhibited a complex change in the DTA signal above 200 °C, which was not observed in non-supported Al₂O₃ and can be attributed to the reduction of Rh precursors (Fig. S15). In contrast, monometallic Pd/Al₂O₃ and Pt/Al₂O₃ showed DTA profiles similar to that of non-supported Al₂O₃, and no additional thermal events attributed to the reduction of metal precursors were observed (Fig. S15). This is likely because the Pd and Pt species had already been reduced at room temperature during the initial H₂ purging step prior to the measurement. When RhPdPt/Al₂O₃ was heated with H₂ supplied from the initial stage of temperature increase, as in the conventional heating method,¹⁶ stepwise changes in the DTA profile were observed above 200 °C (Fig. 2 right, red line), indicating the isolated reduction and crystallization of Rh species. On the other hand, such stepwise changes in the DTA profile were not observed when the introduced gas was changed from H₂ to Ar (Fig. 2 right, green line), and the resulting DTA profile was qualitatively similar to that of non-supported Al₂O₃ recorded under the same conditions (Fig. 2 right, gray line). These experimental results suggested that the alloying mechanism in the gas-switch-triggered reduction method was completely different from that in the conventional mechanism associated with the use and/or formation of multiple complex salts.^19,20 During the temperature increase, Rh, Pd, and Pt precursors supported on γ-Al₂O₃ were stable in Ar, but they were triggered to transform into a nanoalloy when the treatment gas was switched from Ar to H₂ at high temperature.

Fig. 2 TG-DTA profiles of untreated RhPdPt/Al₂O₃ and non-supported Al₂O₃. The samples were heated in increments of 10 °C min⁻¹ to 600 °C under flowing Ar or H₂.

Determination of alloying Rh, Pd, and Pt supported on γ-Al₂O₃

To determine the alloying of Rh, Pd, and Pt supported on γ-Al₂O₃, we used XAS analysis to investigate the changes in the local structure of both RhPdPt/Al₂O₃ prepared by the gas-switch-triggered reduction method and the conventional heating method. Fig. 3 shows the Rh K-edge, Pd K-edge, and Pt L₃-edge Fourier transform (FT) extended X-ray absorption fine structure (EXAFS) spectra of RhPdPt/Al₂O₃ treated by the gas-switch-triggered reduction method and the conventional heating method (labeled as non-switching). For comparison, the FT-EXAFS spectra of the corresponding monometallic catalysts and bulk powder references are also displayed. Fig. S16 and S17 show the XANES spectra and EXAFS oscillations, respectively, before FT. In all edges, the FT-EXAFS spectra of the non-switched RhPdPt/Al₂O₃ were not identical to those of the monometallic catalysts or bulk powder references, but retained some of their spectral characteristics, particularly in the peaks corresponding to pure metal–metal scattering, such as Rh–Rh. These observations suggested that only partial alloying had occurred or that individual metallic domains were retained with interfacial contact rather than being fully mixed at the atomic level. Among the three absorption edges, the Rh K-edge spectrum exhibited the smallest spectral changes from the Rh reference, which may be attributed to the independent reduction and crystallization behavior of the Rh species observed in the TG-DTA. On the other hand, the FT-EXAFS spectra of gas-switched RhPdPt/Al₂O₃ showed more pronounced deviations from the monometallic and reference spectra. The observed split peaks at 2–3 Å have been reported to be a sign of a change in scatter caused by alloy formation.²¹ The implication is that γ-Al₂O₃-supported Rh, Pd, and Pt were alloyed by the gas-switch-triggered reduction method. However, the peaks attributed to scattering by oxygen were also identified at approximately 1.5 Å. In particular, at the Rh K-edge, the peaks attributed to the Rh–O bond were more pronounced than those at the Pd K-edge and Pt L₃-edge. These results suggested that metals alloyed by the gas-switch-triggered reduction method were prone to oxidation by air exposure and that Rh was particularly susceptible. The oxidation of alloyed metals may lead to the restructuring of particles, and the small, independent Rh particles observed in the EDXS map in Fig. 1 (bottom) may have been a consequence of such restructuring. To implement the gas-switch-triggered reduction method, it will be necessary to overcome the problem of atmospheric oxidation of metals by optimizing the preparation conditions and handling of catalyst. However, the gas-switch-triggered reduction method can be merged into the pretreatment process before catalyst evaluation using a continuous gas flow system. In that case, in situ alloy formation would be expected to occur without the oxidation of the alloyed metals because of the absence of air exposure.

Fig. 3 (a) Rh K-edge, (b) Pd K-edge, and (c) Pt L₃-edge FT-EXAFS spectra of RhPdPt/Al₂O₃ treated by the gas-switch-triggered reduction method and the conventional heating method (labeled as non-switching). All the spectra were recorded under atmospheric conditions.

Effect of alloying on the catalytic performance in nitrile hydrogenation under ambient conditions

The effect of alloying Rh, Pd, and Pt on the catalytic performance was evaluated using nitrile hydrogenation as a model reaction because the effect of alloying in the case of that reaction has been reported to be clearly reflected as a difference in catalytic performance.²² Although this reaction is recognized as a cost-effective approach for synthesizing amines and imines essential for organic chemistry (Fig. S18),²³ it generally requires heating and/or pressurization, even with a catalyst.^24–26 However, high-performance catalysts rarely allow reactions to proceed under 1 bar of H₂ at 25 °C (i.e., ambient conditions).^17,22 In addition, nitriles have been reported to be selectively converted to primary amines, secondary imines, and secondary amines over Pd, Rh, and Pt, respectively.^17,22 Based on these reports, we first investigated the catalytic performance of monometallic Rh/Al₂O₃, Pd/Al₂O₃, and Pt/Al₂O₃ for the hydrogenation of benzonitrile under ambient conditions (Table 1). After 1 h of reaction, Rh/Al₂O₃ showed the highest activity among the monometallic catalysts, affording a conversion of 21% and a 15% yield of secondary imine. Pd/Al₂O₃ and Pt/Al₂O₃ exhibited limited reactivity, with conversions of 3% and 2%, respectively. The product selectivities were also consistent with previous reports.^17,22 STEM analysis revealed average particle sizes of 1.9 nm for Rh, 7.7 nm for Pd, and 4.4 nm for Pt (Fig. S19–S21). These trends were consistent with our previous report,¹⁷ in which monometallic Rh, Pd, and Pt catalysts with similar particle sizes exhibited activities in the order of Rh > Pd > Pt for benzonitrile hydrogenation. Moreover, we previously demonstrated that this reaction is structure-sensitive and is significantly promoted over Rh catalysts with smaller particle sizes.^17,23 The relatively high activity of Rh/Al₂O₃ in the present study can be attributed to its small average particle size (1.9 nm). In addition, RhPdPt/Al₂O₃ prepared by the conventional heating method (non-switching) afforded a conversion of 18% and a 16% yield of secondary amine, outperforming the monometallic catalysts despite a relatively large particle size (5.1 nm, Fig. S1). This result suggested that the enhanced activity may originate from the cooperative interaction among the different metal domains, as previously observed in our study on physically mixed Pd and Pt catalyst.²² In contrast, the RhPdPt/Al₂O₃ prepared by the gas-switch-triggered reduction method (gas-switching) exhibited slightly higher activity, achieving a 23% conversion and a 19% yield of secondary amine. Notably, this catalyst had a significantly larger average particle size (14.4 nm, Fig. S2) than both the non-switched and monometallic catalysts. Considering the lower dispersion, the observed performance implied an enhanced intrinsic activity likely due to the formation of random alloys. Although the direct comparison of the XANES features was complicated by differences in the particle size among the catalysts, the observed improvement likely reflects the alloying-induced changes in the electronic and geometric properties of the active sites. Interestingly, when the reaction time was extended to 3 h, the conversion over the non-switched RhPdPt/Al₂O₃ increased markedly to 72%, with a 65% yield of secondary amine, surpassing the gas-switched RhPdPt/Al₂O₃. This reversal in the performance trend may be attributed to the differences in the particle size. As the reaction proceeds toward a higher conversion in a batch system, the local concentration of the substrate near the catalyst surface decreases, and mass transfer becomes increasingly important. Under such diffusion-limited conditions, the smaller particles of the non-switched catalyst (5.1 nm) offer an advantage over the larger particles of the gas-switched catalyst (14.4 nm), enabling more efficient substrate access to the active sites. Then, we tuned the particle size of alloyed RhPdPt supported on γ-Al₂O₃ by changing the Rh precursor from Rh chloride to Rh nitrate and succeeded in further doubling the catalytic performance of the original RhPdPt/Al₂O₃ by downsizing the particle size to 1.7 nm (Fig. S22 and 4). The catalytic activity of the downsized RhPdPt/Al₂O₃ in the hydrogenation of benzonitrile was about 18 times that of monometallic Pt/Al₂O₃ (Fig. 4), which favors the formation of secondary amines like RhPdPt/Al₂O₃. This exceptional performance was close to that of previously reported highest-performing catalysts under the same reaction conditions (Tables S1).²²

Table 1 Benzonitrile hydrogenation to benzylamine (1), benzylidenebenzylamine (2), and dibenzylamine (3) under ambient conditions

Catalyst	Conversion^a (%)	Yield^a (%)			Particle size^b (nm)
		(1)	(2)	(3)
a Conversion and yield were calculated by gas chromatography with flame ionization detection using dodecane as an internal standard. b Particle size was estimated by STEM images. N.S. = non-switching (conventional heating); G.S. = gas-switching (gas-switch-triggered reduction method).
Rh/Al2O3	21	0	15	0	1.9
Pd/Al2O3	3	2	0	1	7.7
Pt/Al2O3	2	0	0	1	4.4
RhPdPt/Al2O3_N.S.	18	1	0	16	5.1
RhPdPt/Al2O3_G.S.	23	2	0	18	14.4

---

Fig. 4 Time courses of benzonitrile conversions and dibenzylamine (secondary amine) yields.

Conclusions

We first developed a simple, impregnation-based method for the alloying of γ-Al₂O₃-supported immiscible Rh, Pd, and Pt, where all metal cations were simultaneously reduced by switching the treatment gas from Ar to H₂ at high temperatures. This method required none of the special equipment or procedures used in previous alloying studies and could also be merged into a pretreatment process before catalyst evaluation, where in situ alloy formation would occur without the oxidation of the alloyed metals because of the absence of air exposure. In this study, the gas-switch-triggered reduction method was used to alloy ternary Rh, Pd, and Pt supported on γ-Al₂O₃, and it succeeded in remarkably improving their catalytic performances in nitrile hydrogenation under ambient conditions. These findings will accelerate design of catalysts based on the differences of their crystalline nature by random alloying and will provide a bridge to their industrial application.

Author contributions

Yoshihide Nishida: designing of research, data collection, and writing – editing. Takaaki Toriyama and Tomokazu Yamamoto: conducting STEM-EDXS analysis. Katsutoshi Sato: conducting X-ray absorption spectroscopy. Katsutoshi Nagaoka: review and editing. Masaaki Haneda: designing of research.

Conflicts of interest

There are no conflicts of interest to declare.

Data availability

Supplementary information on characterization is available. See DOI: https://doi.org/10.1039/D5CY00654F.

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

Acknowledgements

We acknowledge the writing support of Kohei Kusada and Hiroshi Kitagawa of Kyoto University. X-ray absorption analyses were performed at the beamline BL01B1 in SPring-8 (proposal no. 2021B1640, 2022B1920, and 2023B1679) and at BL11S2 in Aichi SR (proposal no. 202502137). This work was supported by JSPS KAKENHI Grant Numbers JP21K20486 and JP23K13604, and by The Naito Science & Engineering Foundation.

References

1. A. Wong, Q. Liu, S. Griffin, A. Nicholls and J. R. Regalbuto, Science, 2017, 358, 1427–1430.
2. K. Murata, Y. Mahara, J. Ohyama, Y. Yamamoto, S. Arai and A. Satsuma, Angew. Chem., Int. Ed., 2017, 56, 15993–15997.
3. K. Era, K. Sato, S.-I. Miyahara, T. Naito, K. De Silva, S. Akrami, H. Yamada, T. Toriyama, T. Tamaoka, T. Yamamoto, Y. Murakami, K. Inazu and K. Nagaoka, Sustainable Energy Fuels, 2024, 8, 2593–2600.
4. K. Sato, H. Higuchi and K. Nagaoka, Chem. Lett., 2024, 53, 1–5.
5. M. Machida, K. Nishiyama, M. Furukubo, H. Yoshida, K. Awaya, J. Ohyama, M. Tsushida, M. Suwa, Y. Nagao, Y. Endo and T. Wakabayashi, ACS Appl. Nano Mater., 2024, 7, 766–776.
6. K. Kusada, H. Kobayashi, R. Ikeda, Y. Kubota, M. Takata, S. Toh, T. Yamamoto, S. Matsumura, N. Sumi, K. Sato, K. Nagaoka and H. Kitagawa, J. Am. Chem. Soc., 2014, 136, 1864–1871.
7. K. Sato, H. Tomonaga, T. Yamamoto, S. Matsumura, N. D. B. Zulkifli, T. Ishimoto, M. Koyama, K. Kusada, H. Kobayashi, H. Kitagawa and K. Nagaoka, Sci. Rep., 2016, 6, 28265.
8. K. Kusada, D. Wu and H. Kitagawa, Chemistry, 2020, 26, 5105–5130.
9. Y. Jiang, Y. Zhu, Y. Jia and S. Dai, CrystEngComm, 2025, 27, 302–306.
10. J. De Bellis, H. Petersen, J. Ternieden, N. Pfänder, C. Weidenthaler and F. Schüth, Angew. Chem., Int. Ed., 2022, 61, e202208016.
11. Z. Chen, T. Liu, H. Zhang, B. Pang, Y. Sun, L. Hu, Q. Luo, X. Liu, L. Cao and T. Yao, Nano Res., 2024, 17, 6032–6037.
12. S. Masuda, K. Mori, T. Sano, K. Miyawaki, W.-H. Chiang and H. Yamashita, ChemCatChem, 2018, 10, 3526–3531.
13. S. Masuda, K. Shun, K. Mori, Y. Kuwahara and H. Yamashita, Chem. Sci., 2020, 11, 4194–4203.
14. K. Shun, K. Mori, S. Masuda, N. Hashimoto, Y. Hinuma, H. Kobayashi and H. Yamashita, Chem. Sci., 2022, 13, 8137–8147.
15. J. Luyten, J. De Keyzer, P. Wollants and C. Creemers, CALPHAD: Comput. Coupling Phase Diagrams Thermochem., 2009, 33, 370–376.
16. K. Kato, D. Deng, Y. Kita, K. Kamata and M. Hara, Catal. Sci. Technol., 2022, 12, 5425–5434.
17. Y. Nishida, Y. Wada, C. Chaudhari, K. Sato and K. Nagaoka, J. Jpn. Pet. Inst., 2019, 62, 220–227.
18. S. R. Bharadwaj, A. S. Kerkar, S. N. Tripathi and S. R. Dharwadkar, J. Less-Common Met., 1991, 169, 167–172.
19. P. E. Plyusnin, Y. V. Shubin and S. V. Korenev, J. Struct. Chem., 2022, 63, 353–377.
20. K. Ding, D. A. Cullen, L. Zhang, Z. Cao, A. D. Roy, I. N. Ivanov and D. Cao, Science, 2018, 362, 560–564.
21. C. Cong, S. Nakayama, S. Maenosono and M. Harada, Ind. Eng. Chem. Res., 2018, 57, 179–190.
22. Y. Nishida, K. Sato, C. Chaudhari, H. Yamada, T. Toriyama, T. Yamamoto, S. Matsumura, S. M. Aspera, H. Nakanishi, M. Haneda and K. Nagaoka, Catal. Sci. Technol., 2022, 12, 4128–4137.
23. Y. Nishida, C. Chaudhari, H. Imatome, K. Sato and K. Nagaoka, Chem. Lett., 2018, 47, 938–940.
24. B. Liu, Y. Li, D. Wang, X. Liu, G. She, P. Zhou, Y. Zhao and Z. Zhang, React. Chem. Eng., 2025, 10, 970–974.
25. J.-Q. Bai, M. Ma, H. Liu, Z. Qian, D. Liu, Y. Zhao, Y. Gao, J. Chen, M. Cai and S. Sun, ACS Catal., 2025, 15, 5086–5102.
26. M. Agüero, D. J. Segobia, N. M. Bertero and A. F. Trasarti, Appl. Catal., A, 2025, 699, 120254.

---

Footnote

† Masaaki Haneda died on May 26, 2024.

---