InNi₃C_0.5@C-derived InNi₃ alloy as a coke-resistant low-temperature catalyst for selective butadiene hydrogenation

Abstract

The development of non-precious metal catalysts to replace Pd-based industrial catalysts for 1,3-butadiene hydrogenation is highly desired. However, the cheap transition metal-based candidates developed to date, including Ni, Cu, Co, and Fe, show lower activity than precious metal catalysts, and typically suffer from deactivation due to coke deposition. In this contribution, an InNi₃ alloy catalyst is designed via depletion of the layers of a carbon-capped interstitial compound InNi₃C_0.5 by two different approaches, hydrogenation and air oxidation, under varying conditions. The carbon depletion processes are followed by combined techniques to determine the full structural and compositional evolution processes, and the resulting catalysts are thoroughly characterized and evaluated for the selective hydrogenation of 1,3-butadiene. The InNi₃ alloy serves as a coke-resistant low-temperature catalyst for 1,3-butadiene hydrogenation, achieving >94% total butene selectivity at >96% conversion at a mild temperature of 318 K. Post characterization of the catalyst after long-term testing verifies the structural robustness and excludes the accumulation of carbonaceous deposits, thus confirming the InNi₃ alloy as a novel and promising system for selective hydrogenations.

---

Introduction

The selective hydrogenation of 1,3-butadiene (BD) is vital to the purification of alkadienes in the upgrading of C4 species to polymers and other valuable chemicals, in which the content of BD residuals should be strictly controlled, as even a trace amount (>200 ppm) will cause catalyst deactivation in the polymerization units.^1–4 In this context, selective hydrogenation is the most effective approach, and depends on the use of efficient catalysts to hydrogenate only one C=C bond of the BD molecule while avoiding over-hydrogenating the other valuable olefins in the C4 stream. To date, precious metals based on Pd,^5–8 Au,^9–11 and Pt (ref. 12–14) have been identified as the most effective catalysts, among which promoter-modified Pd catalysts have been applied in industry. On the other hand, replacing precious metal catalysts with cheap candidates is attractive in heterogeneous catalysis. For BD hydrogenation, distinct catalytic systems based on cheap transition metals have been reported in the last decade, including Cu,^15–18 Ni,^19–21 Fe,²² Co,^23,24 and Mo.^25,26 Nonetheless, catalyst deactivation owing to the deposition of carbonaceous materials is frequently reported in the literature (Cu,^17,18 Ni,²⁰ Fe,²² Co,^23,24 Mo (ref. 25 and 26)), causing a drop in activity and alteration in the product distributions. Furthermore, most of these catalysts are operated at relatively high temperatures (>373 K) in order to achieve a reasonable performance.

Nickel is well-known for diverse hydrogenations, but its applications in BD hydrogenation are still rare. Previous studies reveal that catalyst deactivation and over-hydrogenation, particularly at high conversions, are the key concerns for Ni particles.^19–21 Coating supported Ni catalysts with ionic liquids^19,20 and designing ZnNi₃C_0.7 (ref. 21) interstitial compounds have been proven as effective methods to enhance the performance, which is attributed to the electronic effects induced by the ionic liquids or the segmentation atoms leading to altered adsorptions of the substrates.^19,27 Meanwhile, the potentials of InNi₃C_0.5 in selective hydrogenations (CO₂,^28–30 dimethyl oxalate³¹) have only recently been recognized. Inspired by these promising literature results, herein, for the first time, we report InNi₃ derived from carbon shell-capped InNi₃C_0.5 particles as a highly coke-resistant non-precious metal catalyst for low-temperature BD hydrogenation. The structure and composition evolution of the hybrids during depletion of the carbon shells via both air oxidation and H₂ reduction are unveiled by combined characterization techniques. In contrast to the previous cheap metal catalysts, the InNi₃ alloy featured high resistance to coking and good butene selectivity under mild reaction conditions (>90% butene yield at 318 K).

Results and discussion

Synthesis and characterization of InNi₃C_0.5@C

The bimetallic interstitial compound preserved in carbon shells (InNi₃C_0.5@C) was prepared by a simple one-pot solid-state reaction between melamine, Ni(OH)₂ and In(OH)₃ in H₂ (see the Experimental section). In contrast to the previous approach, which typically involved the preparation of a bimetallic alloy followed by the insertion of interstitial carbon atoms by carbonization with C₂H₂, the developed method is more straightforward and without the use of any solvents. Powder X-ray diffraction patterns (PXRD) reveal that all the diffraction patterns of the resulting product can be perfectly indexed to InNi₃C_0.5 (PDF#28-0468, Fig. S1†). N₂ sorption analysis showed a specific surface area (S_BET) of 60 m² g⁻¹ for the sample (Fig. S2†). The relatively high S_BET is likely because of the deposition of porous carbon shells on the InNi₃C_0.5 crystals. To prove the core–shell structures, we performed high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM, Fig. 1). The high-resolution images in Fig. 1a and b clearly display carbon shells of several nanometres in size, surrounding the InNi₃C_0.5 particles. The elemental color mapping images in Fig. 1c further illustrate the geometrical relevance of Ni and In, and the broader distribution of C around these particles, thus unambiguously pointing to the formation of InNi₃C_0.5@C core–shell structures.

Fig. 1 a and b, High-resolution STEM images of the as-prepared InNi₃C_0.5@C with, c, elemental color mapping results.

Depleting the shells of InNi₃C_0.5@C

To remove the carbon shells, two different methods were adopted: air oxidation and methanation in H₂. The derived solids were denoted as AX or HX, respectively, where X represents the applied temperature in K. PXRD revealed the typical facets of InNi₃C_0.5 for HX and A573, whereas mixed In₂O₃ and NiO were detected for A1023 (Fig. 2a). Nonetheless, the effectiveness of H₂ reduction for carbon removal was confirmed by STEM observation of H1023, showing clearly reduced carbon layers surrounding the metal particles (Fig. S3†). In situ PXRD was further performed to track the phase evolution of InNi₃C_0.5@C during air oxidation (Fig. 2b). In this case, the two most intense diffraction peaks gradually weakened in intensity and shifted to lower 2θ with increasing temperature. Formation of In₃Ni₂ was clearly evidenced at 823 K, as well as the appearance of In₂O₃ and InNi₃. When further increasing the temperature to 1073 K, the intensities of In₃Ni₂ decreased while In₂O₃ and InNi₃ predominated, highlighting the different thermal stabilities. The above PXRD results confirmed the more aggressive destruction of InNi₃C_0.5@C by air oxidation than H₂ reduction, which was corroborated by Raman spectroscopy (Fig. 2c). The typical D and G bands at 1350 and 1580 cm⁻¹, respectively, were found for HX and A573, suggesting still the presence of carbonaceous residuals on these samples. Meanwhile, additional bands at 570 cm⁻¹ likely relating to Ni–In alloys appeared for A573 and H1023, due to the partial depletion of the shells. In contrast, those carbon-defect related bands were not detected for A1023, and new bands at 518, 1069, and 1460 cm⁻¹ associated with In₂O₃ (ref. 32) and NiO (ref. 33) appeared, suggesting the complete removal of the carbon shells. These results were further corroborated by X-ray photoelectron spectroscopy (XPS). Elemental analysis showed a significant increase of surface metal species for A1023 as compared to InNi₃C_0.5@C (4.44 vs. 1.21% for Ni, and 5.53 vs. 1.08% for In), plus the disappearance of N signals (Fig. S4†). Additionally, the Ni⁰ and In⁰ species, typically observed for InNi₃C_0.5 in previous studies,^29,30 also disappeared after the oxidation treatment (Fig. S5†). Thermogravimetric analysis (TGA) was performed to quantify the carbon residuals (Fig. 2d). The TGA profiles differed greatly from each other, and showed sequential weight loss and weight gain due to the complicated carbon combustion and oxidation of the metallic components. By supposing that InNi₃C_0.5@C was completely oxidized into In₂O₃, NiO, and CO₂, the carbon residuals were calculated to be 21%, 20%, 15%, 10%, and 0, for InNi₃C_0.5@C, H573, H1023, A573, and A1023, respectively (see the Experimental section for details).

Fig. 2 a, PXRD; b, in situ PXRD of InNi₃C_0.5@C in air; c, Raman spectra; d, TGA profiles. STEM images of e, InNi₃C_0.5@C and f–h, A1023 after H₂ reduction. The inset in e shows the magnified image.

Since air oxidation at 1023 K can completely remove the carbon shells, the compositions of A1023 after H₂ reduction at 673 K were studied. STEM was used to visualize the aggregated particles without any carbon residuals, and the lattice fringes of two alloy phases (In₃Ni₂ and InNi₃) were verified (Fig. 2e–h), agreeing well with the PXRD result (Fig. 2a). Combining all the above characterization results, the full structure and composition evolution of InNi₃C_0.5@C can be depicted as in Fig. 3. Air oxidation removes the carbon shells. The interstitial composition can be preserved in H₂ reduction but gradual phase transformation from InNi₃C_0.5 to a mixed Ni–In alloy and finally to mixed oxides occurred during oxidation. The alloys can be easily restored by subsequent reduction of the mixed oxides.

Fig. 3 Schematic illustrations of the structure and composition evolution of InNi₃C_0.5@C subject to different treatments.

Catalytic performance in BD hydrogenation

Pristine InNi₃C_0.5@C and the H₂- or air-treated sample were applied for the first time to BD hydrogenation in a fixed-bed reactor. The impact of the H₂ : BD ratio was first evaluated in a temperature-ramping study on InNi₃C_0.5@C, showing a shift of the activity curves to lower temperatures and slightly reduced selectivity to total butene when the ratio was increased from 20 to 50, and no significant change at 100 (Fig. 4a and b). Further comparisons of the performances of the developed catalysts were conducted at H₂ : BD = 50 (Fig. 4c and d). Divergent catalytic behaviors were observed for the catalysts subject to different treatments. Namely, the air-oxidized samples showed much higher activity and slightly improved selectivity to total butene as compared with the pristine InNi₃C_0.5@C. For example, the temperature corresponding to 96% BD conversion was markedly lower, by 75 K (from 423 to 348 K). In contrast, the H₂-treated samples showed lower activity in comparison to InNi₃C_0.5@C, although the total butene selectivity was enhanced. These different catalytic responses can be understood by considering the different structures and compositions of the catalysts as revealed in the previous sections. The activities of the four catalysts with InNi₃C_0.5 particles preserved by carbon shells followed the order A573 > InNi₃C_0.5@C > H573 > H1023, hinting at likely interplays between the carbon residuals and the particle sizes (e.g., the greatly sharpened PXRD diffractions of H1023 suggested enlarged particle sizes of InNi₃C_0.5). A1023 displayed a higher activity, even though N₂ sorption isotherms revealed much lower specific surface areas of A1023 than InNi₃C_0.5@C (11 vs. 60 m² g⁻¹, Fig. S2†), thus indicating reasonable hydrogenation potentials of the Ni–In alloys. Temperature-programmed desorption of H₂ (H₂-TPD) was further performed to understand the catalytic differences (Fig. S6†). A1023 without the carbon shells showed three desorption peaks at 433, 801, and 1107 K, whereas the pristine InNi₃C_0.5@C only displayed a low-temperature desorption peak at 380 K. Therefore, the stronger H₂ adsorption propensity of A1073 agrees well with the better hydrogenation activity. Given the superior catalytic performance of A1023, the long-term stability was evaluated. Negligible deactivation was observed under the high conversion conditions (X_BD > 98%) and the total butene selectivity was maintained at >87% (Fig. 4e). The catalyst stability was also verified at lower conversions (ca. 80%) by increasing the gas hourly space velocity (GHSV) from 90 000 to 300 000 cm³ g_cat⁻¹ h⁻¹. The catalyst remained essentially stable in the 40 h test (Fig. S7†), although a slow activity drop in the first 16 h was observed, which might be related to the phase transition from In₃Ni₂ to InNi₃ (vide infra). The GHSV was further varied and near full conversion was reached at 318 K and 30 000 cm³ g_cat⁻¹ h⁻¹ (Fig. S8†). To assess the catalyst low-temperature potential, the stability at 318 K was evaluated (Fig. 4f). After a very small activity drop in the first 12 h accompanied by an increase in butene selectivity, a stable performance was achieved in the next 24 h with >94% total selectivity to butene at >96% BD conversion.

Fig. 4 The catalytic performance of a and b, pristine, and c and d, partially carbon shell-removed InNi₃C_0.5@C in BD hydrogenation. e and f Time-on-stream performance of A1023. Reaction conditions: GHSV = 90 000 cm³ g_cat⁻¹ h⁻¹, 0.69 vol% BD balanced in N₂, a and b, H₂ : BD = 20–100, c–e, H₂ : BD = 50, e, T = 353 K, f, 318 K, H₂ : BD = 100, GHSV = 30 000 cm³ g_cat⁻¹ h⁻¹.

Characterization of the used catalysts and kinetics

To probe the properties of the used catalyst, A1023 after the 24 h stability test was characterized using different techniques. PXRD revealed that all the diffraction lines can be indexed to InNi₃, and the other alloy, In₃Ni₂, in the reduced sample was not observed (Fig. 5a), thus supporting our earlier speculation on the activity drop. TEM observations evidenced the similar structures of the interconnected nanoparticles, as well as the distinct lattice fringes of the (101) and (201) facets of the InNi₃ alloys (Fig. S9†). Additionally, coke deposition was excluded by TGA and Raman spectra (Fig. 5b and c), showing weight gain of 21.1%, close to the theoretical values (24.7% for InNi₃ and 22.5% for In₃Ni₂), and the absence of Raman features for carbonaceous materials. Similar characterization results were acquired for the catalyst evaluated at 318 K for 36 h (Fig. S10†), except that both InNi₃ and In₃Ni₂ were preserved, hinting that the phase transition was much favored at high temperatures.

Fig. 5 Characterization of A1023 after the 24 h evaluation in BD hydrogenation: a, PXRD; b, TGA; c, Raman spectra.

To shed light on the reaction mechanism, the reaction kinetics were further studied. By varying the partial pressures of the respective reactants while keeping the same total flow rate, the partial reaction orders of H₂ and BD on A1023 were determined to be 1.1 and −1.2, respectively (Fig. 6). The negative BD order suggests the significantly stronger adsorption of the alkadiene on the alloy surface than H₂. This trend agrees well with those observed for Pd- and Pt-based catalysts.^8,34 Therefore, it is likely that the hydrogenation on the InNi₃ alloy also follows the classic Horiuti–Polanyi mechanism for alkyne hydrogenation on Pd-based catalysts.³⁵ Since BD hydrogenation is a chain reaction, the timely desorption of butene is crucial in order to achieve high selectivity, as well as to avoid alkene oligomerization, which is suspected to induce coke accumulation. In this study, the main efforts were placed on understanding the impacts of the carbon shell depletion of the interstitial compounds. A joint study coupling the synthesis of structurally well-defined catalysts and density functional theory calculations is expected to produce more fundamental insights into the nature of the active sites and reaction mechanisms.

Fig. 6 The partial reaction orders of the respective reactants in BD hydrogenation on A1023.

Conclusions

In summary, we have reported an InNi₃ alloy as a new and stable non-precious metal catalyst for the selective hydrogenation of 1,3-butadiene. This alloy was derived from a bimetallic interstitial compound, InNi₃C_0.5, preserved in a carbon shell. With the aim of removing the carbon shell, the evolution of the structure and composition during air oxidation and H₂ reduction was studied in detail. Complete depletion of the shell was achieved by oxidation at 1023 K, accompanied by the formation of mixed metal oxides, which can be converted to In–Ni alloys by H₂ reduction. When evaluated in BD hydrogenation, a remarkable performance was achieved at a low temperature of 318 K, with >94% selectivity to total butene at >96% BD conversion. In contrast to the other known non-precious metal catalysts that frequently suffer from quick deactivation, InNi₃ showed high coke resistance in the hydrogenation reaction, thus offering relatively stable operation.

Experimental

Materials

Melamine (C₃H₆N₆, 99%) was purchased from Aladdin. Ni(OH)₂ (Ni 61%) and In(OH)₃ (99.99%) were purchased from Macklin. All reagents were obtained from commercial suppliers without further purification prior to use.

Catalyst synthesis

InNi₃C_0.5@C: the bimetallic interstitial compound InNi₃C_0.5 preserved in carbon shells (InNi₃C_0.5@C) was prepared by a one-pot solid-state synthetic approach. Typically, a mixture of commercial nickel hydroxide and indium hydroxide was used as the metal precursors, with melamine as the carbon precursor, where the mass ratio of the metal precursors to melamine was fixed at 1 : 5. By mechanical mixing, 0.3735 g of In(OH)₃, 0.6265 g of Ni(OH)₂, and 5 g of melamine were ground in a mortar for 15 min until uniformly mixed. The obtained greenish powders were placed in a tubular furnace under a flow of H₂ (60 mL min⁻¹) and annealed at 973 K (duration 4 h, ramp rate 5 K min⁻¹). After cooling to room temperature, the solids were stored in air.

Decarbonized InNi₃C_0.5@C: to remove the carbon shells, the as-prepared InNi₃C_0.5@C solids were placed in a tubular furnace and treated by two different methods. The first method was oxidation in static air for at 573 and 1023 K (duration 3 h, ramp rate 5 K min⁻¹). The derived samples were denoted by A573 and A1023. The other method was methanation via H₂ reduction. The solids were treated in flowing H₂ (30 mL min⁻¹) at 573 and 1023 K (duration 3 h, ramp rate 5 K min⁻¹). The obtained materials were denoted by H573 and H1023.

Characterization methods

N₂ sorption was performed at 77 K on a Quanta chrome NT3LX-2 instrument. Prior to the analysis, the powder samples were degassed at 573 K for 3 h. The specific surface areas (S_BET) were calculated by using the Brunauer–Emmett–Teller (BET) method. Powder X-ray diffraction (PXRD) patterns were conducted on an X'Pert 3, PANalytical X-ray diffractometer using Cu K_α radiation in a scanning angle (2θ) range of 20–80° at a speed of 2° min⁻¹. In situ PXRD patterns were recorded with a Rigaku SmartLab-9 kW D/teX Ultra250 diffractometer using a Cu K_α radiation source operated at 40 kV and 200 mA. The powder samples were spread into a high-temperature chamber and the diffraction patterns were recorded at 2θ of 30–90° with a scanning rate of 10° min⁻¹ at room temperature. Then, the samples were heated to 1073 K successively at a rate of 10 K min⁻¹ under an air flow of 30 mL min⁻¹. Thermogravimetric analyses (TGA) of the catalysts were carried out using the NETZSCH STA 449 F5 system with a heating rate of 10 K min⁻¹ (from room temperature to 1073 K) under flowing air. The samples were placed in an α-Al₂O₃ crucible. Raman spectra were acquired on a confocal laser micro-Raman spectrometer (HORIBA, Lab RAM HR Evolution) with a wavelength of 532 nm, 60 s transits per sample, and a spectral resolution of 1 cm⁻¹. Transmission electron microscopy (TEM) was conducted on a FEI Talos F200S. The high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) measurements were carried out on a JEOL-ARM200F electron microscope. The sample powders were dispersed in ethanol by ultrasonication and then the species were obtained by dropping a droplet suspension on a carbon film supported on a copper grid for the analysis. Temperature-programmed desorption of H₂ (H₂-TPD) was conducted on a Micromeritics Autochem II 2920 chemisorber equipped with a thermal conductivity detector and mass spectrometer. Typically, the sample (100 mg) was pretreated in a quartz U-tube reactor with a He gas stream (50 cm³ min⁻¹) at 473 K for 1 h. After cooling to room temperature, a 10% H₂/Ar flow (cm³ min⁻¹) was introduced and the sample was heated to 1273 K at a rate of 10 K min⁻¹. The X-ray photoelectron spectroscopy (XPS) measurements were carried out on a Thermo ESCALAB 250Xi spectrometer (15 kV Al Kα X-ray source). The C 1s peaks were calibrated at the binding energy of 284.8 eV.

Catalytic testing

Butadiene hydrogenation was evaluated in a home-made fixed-bed quartz reactor (4 mm internal diameter, 400 mm length) at ambient pressure. Pellet catalysts of 20–40 mesh in size were used for the catalytic evaluation. Typically, the catalysts were packed between glass wool in the middle of the reactor, which was vertically placed in an electronic oven equipped with a temperature controller. The catalysts were first reduced by H₂ (50 cm³ min⁻¹) for 1 h. After cooling to room temperature, the reactive gases composed of 4.03 vol% BD/N₂, 10 vol% H₂/N₂, and balancing N₂ were introduced into the reactor. The tail gas from the outlet of the reactor was periodically analyzed by using an Agilent 7890B chromatograph equipped with a CEC-Alumina column (30 m × 0.53 mm × 10 μm) and a flame ionization detector. For the kinetic measurements, the BD conversions were restricted to <25% in order to minimize the gradient of the reactants in the feed. The reaction rates were measured at 323 K by varying the relative partial pressures of the respective reactants (P_H2 = 0.17–0.52, P_BD = 0.0086–0.014).

The BD conversion (X_BD) and the product selectivity (S_i) were determined by eqn (1) and (2), respectively:

X_BD (%) = (F_BD,in − F_BD,out)/F_BD,in × 100 (1)

where F_BD,in and F_BD,out are the BD flow rates (mol min⁻¹) from the inlet and outlet, respectively, and F_i,out and F_j,out are the outlet flow rates (mol min⁻¹) of the reaction products. i (j) = 1-butene, trans-2-butene, cis-2-butene, and butane.

The carbon balances (ε, %), estimated by eqn (3), were typically within 100 ± 2 for all the tests:

Estimation of the carbon residual content of InNi₃C_0.5@C

i) Supposing that the content of the carbon shell (not the total carbon content) is x, then the share of InNi₃C_0.5 is 1 − x.

ii) According to the PXRD of the completely oxidized samples (see Fig. 2a), the pristine sample is finally converted into NiO and In₂O₃. Therefore, the chemical reaction during the combustion of InNi₃C_0.5 can be expressed by eqn (4):

InNi₃C_0.5 + 2.75O₂ → 0.5In₂O₃ + 3NiO + 0.5CO₂ (4)

Based on eqn (4), the weight gain y is 20.2%, determined by eqn (5):

y = (4.5 × M_w(O) − 1.2 × M_w(C))/M_w(InNi₃C_0.5) (5)

The final weight change (z) in the TGA is shown in Fig. 2d.

Then z can be determined by eqn (6):

z = (1 − x) × y − x (6)

Therefore:

x = 0.202 − z/1.202.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

We thank the National Natural Science Foundation of China (22372150), Jinhua Science and Technology Plan Project (2022-1-078), National Key Research and Development Project of China (2022YFB3805600), and Zhejiang Normal University for providing financial support (ZZ323205020521005039, KYJ51020910, and YS304320036).

References

1. C. Louis and L. Delannoy, Adv. Catal., 2019, 1–88,  DOI:10.1016/bs.acat.2019.08.002.
2. P. R. Selvakannan, L. Hoang, V. V. Kumar, D. Dumbre, D. Jampaiah, J. Das and S. K. Bhargava, in Catalysis for clean energy and environmental sustainability: Petrochemicals and refining processes, ed. K. K. Pant, S. K. Gupta and E. Ahmad, Springer International Publishing, Cham, 2021, vol. 2, pp. 205–228.
3. M. Wang, Y. Wang, X. Mou, R. Lin and Y. Ding, Chin. J. Catal., 2022, 43, 1017–1041.
4. Q. C. Wei, Y. Chen, Z. Wang, D. Z. Yu, W. H. Wang, J. Q. Li, L. H. Chen, Y. Li and B. L. Su, Angew. Chem., 2022, 61, e202210573.
5. H. Yi, H. Du, Y. Hu, H. Yan, H.-L. Jiang and J. Lu, ACS Catal., 2015, 5, 2735–2739.
6. E. Castillejos-Lopez, G. Agostini, M. Di Michel, A. Iglesias-Juez and B. Bachiller-Baeza, ACS Catal., 2017, 7, 796–811.
7. H. Yan, H. Cheng, H. Yi, Y. Lin, T. Yao, C. Wang, J. Li, S. Wei and J. Lu, J. Am. Chem. Soc., 2015, 137, 10484–10487.
8. Y. Wang, Q. Yuan, L. Shi, Z. Chen, X. Mou, M. Li, J. Dai, X. Song, R. Lin and Y. Ding, Mol. Catal., 2023, 546, 113278.
9. Z. P. Liu, C. M. Wang and K. N. Fan, Angew. Chem., Int. Ed., 2006, 45, 6865–6868.
10. G. C. Bond, Gold Bull., 2016, 49, 53–61.
11. Y. Wang, F. Zhang, M. Wang, X. Mou, S. Liu, Z. Jiang, W. Liu, R. Lin and Y. Ding, Angew. Chem., Int. Ed., 2022, 61, e202214166.
12. Y. Wang, M. Wang, X. Mou, S. Wang, X. Jiang, Z. Chen, Z. Jiang, R. Lin and Y. Ding, Nanoscale, 2022, 14, 10506–10513.
13. Z. Zhang, Y. Zhu, H. Asakura, B. Zhang, J. Zhang, M. Zhou, Y. Han, T. Tanaka, A. Wang, T. Zhang and N. Yan, Nat. Commun., 2017, 8, 16100.
14. K. Yang and B. Yang, J. Phys. Chem. C, 2018, 122, 10883–10891.
15. N. Hu, X.-Y. Li, S.-M. Liu, Z. Wang, X.-K. He, Y.-X. Hou, Y.-X. Wang, Z. Deng, L.-H. Chen and B.-L. Su, Chin. J. Catal., 2020, 41, 1081–1090.
16. G. Totarella, R. Beerthuis, N. Masoud, C. Louis, L. Delannoy and P. E. de Jongh, J. Phys. Chem. C, 2021, 125, 366–375.
17. A. Jalal, Y. Zhao and A. Uzun, Ind. Eng. Chem. Res., 2022, 61, 2068–2080.
18. G. Totarella, J. W. de Rijk, L. Delannoy and P. E. de Jongh, ChemCatChem, 2022, 14, e202200348.
19. A. Jalal and A. Uzun, J. Catal., 2017, 350, 86–96.
20. A. Jalal and A. Uzun, Appl. Catal., A, 2018, 562, 321–326.
21. S. Dong, Y. Pu, Y. Niu, L. Zhang, Y. Wang and B. Zhang, Acta Phys.-Chim. Sin., 2023, 39, 2301012.
22. Z. Wang, L. Santander de Soto, C. Méthivier, S. Casale, C. Louis and L. Delannoy, Chem. Commun., 2021, 57, 7031–7034.
23. Y. Jiang, G. Liu, S. Wu, X. Zhang, X. Dai, Q. Sheng, Y. Wang and H. Wang, Micropor. Mesopor. Mat., 2019, 288, 109557.
24. Z. Chen, L. Guo, Y. Wang, X. Mou, Z. Chen, R. Lin and Y. Ding, Catal. Sci. Technol., 2023, 13, 968–974.
25. R. Hou, K. Chang, J. G. Chen and T. Wang, Top. Catal., 2015, 58, 240–246.
26. Q. Yang, R. Qiu, X. Ma, R. Hou and K. Sun, Catal. Sci. Technol., 2020, 10, 3670–3680.
27. C. S. Spanjers, J. T. Held, M. J. Jones, D. D. Stanley, R. S. Sim, M. J. Janik and R. M. Rioux, J. Catal., 2014, 316, 164–173.
28. C. Meng, G. Zhao, X.-R. Shi, Q. Nie, Y. Liu and Y. Lu, Appl. Catal., B, 2022, 316, 121699.
29. P. Chen, G. Zhao, X.-R. Shi, J. Zhu, J. Ding and Y. Lu, iScience, 2019, 17, 315–324.
30. C. Meng, G. Zhao, X. R. Shi, P. Chen, Y. Liu and Y. Lu, Sci. Adv., 2021, 7, 34348903.
31. J. Zhu, G. Zhao, C. Meng, P. Chen, X.-R. Shi and Y. Lu, Chem. Eng. J., 2021, 426, 130857.
32. R. Cuscó, T. Yamaguchi, E. Kluth, R. Goldhahn and M. Feneberg, Appl. Phys. Lett., 2022, 121, 062106.
33. J. Qiu, T. H. Nguyen, Y. J. Lee, S. Kim, S. Kim, S.-J. Kim, M.-T. Song, W.-J. Huang, X.-B. Chen and I.-S. Yang, Spectrochim. Acta A Mol. Biomol., 2023, 297, 122700.
34. C. Yoon, M. X. Yang and G. A. Somorjai, Catal. Lett., 1997, 46, 37–41.
35. I. Horiuti and M. Polanyi, Trans. Faraday Soc., 1934, 30, 1164–1172.

---

Footnotes

† Electronic supplementary information (ESI) available: Methods of catalyst synthesis, characterization, and testing, plus additional characterization results. See DOI: https://doi.org/10.1039/d3cy01260c

‡ Equal contribution.

---