Pt–Ag and Pt–Ag–Sn bimetallic and trimetallic catalysts supported on γ-Al₂O₃ for direct propane dehydrogenation

Abstract

The synthesis, characterization and catalytic performance assessment of Pt–Ag and Pt–Ag–Sn bimetallic and trimetallic catalysts supported on KOH-treated γ-Al₂O₃ for the direct propane dehydrogenation reaction have been carried out. The bimetallic catalysts contained 0.3 wt% Pt and 0.5 wt% K, with x = 0.1, 0.3, 0.5, 0.7, 0.9, and 1.1, corresponding to the amount of Ag (0.3Pt–xAg/0.5K–Al₂O₃). The trimetallic catalysts contained 0.3 wt% Pt, 0.7 wt% Sn and 0.5 wt% K, with the same x values for Ag (0.3Pt–xAg–0.7Sn/0.5K–Al₂O₃). The catalytic tests were run on real samples (1.8 mm spheres). The 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ composition resulted in the highest yield of propylene (31.4%), highest propylene selectivity (79.3%) and minimum deactivation after 200 min on stream at a gas hourly space velocity of 10 000 cm³ g per catalyst per h at a reaction temperature of 580 °C, comparable to the performance of a DeH-16 commercial catalyst under the same operating conditions. Based on X-ray diffraction, H₂-temperature-programmed reduction, X-ray photoelectron spectroscopy, NH₃-temperature-programmed desorption, electron microscopy, Raman spectroscopy and temperature-programmed oxidation analysis, the performance of the optimum sample was attributed to the synergetic electronic and geometrical effects induced by the formation of Pt–Ag and Pt–Sn alloys and partial coverage of support acidic sites by Ag and Sn clusters.

---

Introduction

Industrial propylene production currently includes three main routes: co-production, by-production and on-purpose production. Direct hydrogenation and oxy-dehydrogenation constitute the main on-purpose production processes. Direct propane dehydrogenation (direct PDH) is still the dominant industrial technology (Catofin and Oleflex processes), and the development of new generations of industrial catalysts is routine. Direct PDH uses Pt and CrO_x-based catalysts, and as far as eco-friendliness and stability are concerned, the first is preferred.¹ High catalyst activity, propylene selectivity and stability against coking are the main challenges in designing good direct PDH catalysts.² Alumina phases (mostly θ, δ and γ) constitute the main support for the active metal of commercial catalysts.³ Extensive research has been carried out over the last two decades on the improvement of the support and selection of metal additives to enhance catalyst performance. Natural clay minerals such as sepiolite,⁴ bentonite⁴ and hydrotalcite⁵ have been introduced as novel supports for the PDH reaction. Furthermore, it has been pointed out that supports should be tuned in terms of their acid/base character and morphology to suppress coke formation at the associated high reaction temperatures.⁶ Regarding the metal promoter, Sn and Zn have been reported to substantially enhance selectivity to propylene and stability against coking due to electronic⁷ and geometrical effects.^8,9 However, different promoters like Ag,¹⁰ Au,¹⁰ Co,¹¹ Cu,^10,12–14 In,^15,16 Mn,¹⁷ Ni (ref. 18) and Ga (ref. 19) have also been subject to investigation, and In has emerged as a potential candidate alongside Sn and Zn.

The main scope of the present study is the investigation of Ag as a promoter for the Pt metal catalyst on γ-Al₂O₃ supports for the direct PDH reaction. Theoretical and experimental studies show that Ag and Pt metals strongly interact electronically and tend to form intermetallic compounds.^20,21 The only work considering Ag as a Pt promoter for the PDH reaction is that of Ren et al.,¹⁰ who investigated Pt–Ag bimetallic catalysts supported on MgAl₂O₄ rather than on a γ-Al₂O₃ support. Their work considered only 0.1 wt% Pt and 0.1 wt% Ag, and was limited to a bimetallic catalyst. It should be stated that Ag was known as an important industrial hydrogenation/dehydrogenation catalyst promoter more than 70 years, and Dow Chemical Co. issued a patent on Ag-promoted selective acetylene hydrogenation catalysts in 1954. Therefore, a comprehensive study of Pt–Ag/γ-Al₂O₃ catalysts is required. In this study, we synthesized a wide range of bimetallic Pt–Ag/γ-Al₂O₃ catalysts with Ag contents of 0.1, 0.3, 0.5, 0.7, 0.9 and 1.1 wt% on a γ-Al₂O₃ support with a constant Pt content of 0.3 wt%. The promoting effect of Ag on Pt–Sn catalysts has not previously been elaborated. The present study also considers the addition of Ag, as a promoter, to Pt–Sn catalysts. A series of Pt–Ag–Sn/γ-Al₂O₃ catalysts have been synthesized with Ag contents of 0.1, 0.3, 0.5, 0.7, 0.9 and 1.1 wt% with constant Pt and Sn contents of 0.3 and 0.7 wt%, respectively. To the knowledge of the authors of this work, no report on trimetallic Pt–Ag–Sn on γ-Al₂O₃ supported catalysts used in the PDH reaction exists. The catalytic performances of the bimetallic and trimetallic catalysts for the direct PDH reaction have been thoroughly studied. Interestingly, it will be shown that the promotional effect of Ag is enhanced in the presence of Sn. A KOH-treated γ-Al₂O₃ support is used throughout this work to suppress coking.

Experimental

Catalyst synthesis

Spherical γ-Al₂O₃ particles (Puralox, Sasol) with an average diameter of 1.8 mm were used as a support. The support was initially impregnated with a KOH (Merck) solution in a rotary evaporator under vacuum at 70 °C for 20 min, followed by overnight drying at 90 °C and calcination in an electrical furnace at 450 °C for 3 h at an initial heating rate of 3 °C min⁻¹. The final support contained 0.5 wt% elemental K and is referred to as 0.5K–Al₂O₃. This treatment has been optimized in our laboratory to minimize the activity of the support towards coke formation at the high PDH reaction temperature employed.²²

Addition of the Pt element was performed by impregnating the 0.5K–Al₂O₃ support with an aqueous solution of hexachloroplatinic acid using H₂PtCl₆·6H₂O (Merck) as a precursor. The impregnation, drying and calcination procedures were similar to the 0.5K–Al₂O₃ support preparation. The final material contained 0.3 wt% Pt and is referred to as 0.3Pt/0.5K–Al₂O₃.

Addition of the Ag element was performed by impregnating the 0.3Pt/0.5K–Al₂O₃ product with an aqueous solution of AgNO₃ (Merck). The impregnation, drying and calcination procedures were similar to the preparation of 0.5K–Al₂O₃ support. The final material is denoted as 0.3Pt–xAg/0.5K–Al₂O₃, where x indicates the amount of elemental Ag in the final product and assumes values of 0.1, 0.3, 0.5, 0.7, 0.9 and 1.1.

Addition of the Sn element was performed by impregnating the 0.3Pt–xAg/0.5K–Al₂O₃ products with a solution of SnCl₂·2H₂O (Merck) in methanol (Dr. Mojallali Chemical Industries Complex, Iran). The impregnation, drying and calcination procedures were similar to the 0.5K–Al₂O₃ support preparation. The final material contained 0.7 wt% elemental Sn and is accordingly denoted as 0.3Pt–xAg–0.7Sn/0.5K–Al₂O₃. In this case, x = 0 was also used (0.3Pt–0.7Sn/0.5K–Al₂O₃).

Notably, using the order Pt, Sn, Ag was not appropriate. As the Sn source material (SnCl₂) contains Cl atoms, the addition of AgNO₃ results in the immediate formation of irregular AgCl particles with a very broad size distribution. Most of the AgCl particles formed detach and enter the liquid phase. This preparation method was discarded because it was not reproducible.

Catalyst characterization

Phase analysis was performed using an EQUINOX 3000 (Inel, France) instrument (40 kV, 30 mA, λ = 1.5406 Å, CuKα radiation source). X-ray photoelectron spectroscopy (XPS) was carried out using a Bes Tek (Germany) instrument (10⁻¹⁰ mbar, AlKα radiation source with an energy of 1486.6 eV). The experimental error in XPS data analysis was below 0.2 eV. Field emission scanning electron microscopy (FE-SEM) was performed using a MIRA 3 LMU (TESCAN, Czech Republic) instrument. Energy-dispersive X-ray spectroscopy (EDX) was performed using the same instrument. Nitrogen adsorption/desorption isotherms of the catalyst samples were obtained using a Micromeritics TriStar II apparatus. Raman spectroscopy of the spent catalyst samples was performed using a Takram P50C0R10 Raman system with an excitation wavelength of 532 nm.

Hydrogen temperature-programmed reduction (H₂-TPR) of the catalyst samples was performed using a Micromeritics ChemiSorb 2750 instrument (outgassing at 350 °C for 2 h under an Ar flow of 20 cm³ min⁻¹, and reduction by a gas mixture of 5 vol% H₂ in Ar at a flow rate of 20 cm³ min⁻¹ with a temperature ramp of 10 °C min⁻¹).

Ammonia temperature-programmed desorption (NH₃-TPD) was carried out using a Micromeritics ChemiSorb 2750 apparatus (outgassing at 350 °C for 60 min under a He flow of 20 cm³ min⁻¹, exposure to ammonia using a gas mixture of 5 vol% NH₃ in He at a flow rate of 30 cm³ min⁻¹ at room temperature, followed by He purging at a flow rate of 20 cm³ min⁻¹ for 60 min, and heat ramping at a rate of 10 °C min⁻¹ up to 900 °C).

Temperature-programmed oxidation (TPO) of the spent catalysts was carried out with the same instrument used for H₂-TPR analysis, where the samples were heated to 800 °C with a constant heating rate of 10 °C min⁻¹ in a controlled atmosphere of 5 vol% O₂ in He with a constant flow rate of 30 cm³ min⁻¹.

Catalytic performance assessment

A fixed-bed quartz reactor with an ID of 10 mm under atmospheric pressure was used. After charging, the catalyst was first reduced at 400 °C for 120 min under a H₂ flow of 30 cm³ min⁻¹. Then, the hydrogen flow was stopped, and a nitrogen flow of 40 cm³ min⁻¹ was introduced, while the system was heated to the reaction temperature (580 °C) at a heating rate of 4.5 °C min⁻¹. Then, a flow of 30 cm³ min⁻¹ of H₂ and 30 cm³ min⁻¹ of propane was added to the nitrogen flow to provide the required feed composition and total flow (100 cm³ min⁻¹). The applied gas hourly space velocity (GHSV) was 10 000 cm³ g per catalyst per h. Feed and product streams were analyzed by a YL 6500 gas chromatograph apparatus equipped with an FID detector and an HP-PLOT/Q capillary column (ID = 0.53 mm, length = 30 m).

The propane conversion (X(t)), propylene selectivity (S(t)) and propylene yield (Y(t)) at any time (t) were calculated as below:

Y(t) = X(t)S(t) (3)

where x_iin is the molar fraction of component i in the feed, and x_iout is the molar fraction of component i in the outlet gas stream. In the above equations, the presence of low concentrations of methane and ethane impurities in the feed gas have been accounted for.

Catalyst deactivation was modelled using first order deactivation kinetics:

a(t) = a₀e^−k_dt (4)

where a(t) is the catalyst activity as a function of time, a₀ is the initial activity and k_d is the deactivation constant. The value of k_d was calculated using the formula given by Gao et al.:²³

Results and discussion

XRD

Fig. 1 shows the XRD patterns of the fresh samples 0.3Pt/0.5K–Al₂O₃, 0.3Pt–0.9Ag/0.5K–Al₂O₃ and 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ and the used sample (200 min on stream) 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃. The typical peaks of γ-Al₂O₃ are indicated (2 theta = 19.6°, 37.6°, 39.5°, 45.8° and 66.8°) and appear with negligible relative intensity change in all the samples. Upon addition of the Ag component, three characteristic peaks of Ag₂O appear at 26.5°, 32.6° and 76.14°, belonging to the (110), (111) and (311) planes. The position and relative intensity of the (110) and (111) peaks are in accordance with JCPDS 42-1104. The integrated intensities of the Ag₂O characteristic peaks are significantly higher for the 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ sample. This is while the latter sample contains the same relative content of Ag as the 0.3Pt–0.9Ag/0.5K–Al₂O₃ sample. This may be explained based on the size of the Ag₂O clusters. The addition of Sn results in larger Ag₂O particles, eventually leading to stronger Ag₂O peaks. Ag₂O nanoparticles in the range of 1–2 nm are known to exhibit distinct peaks.²⁴

Fig. 1 XRD spectra of the 0.3Pt/0.5K–Al₂O₃, 0.3Pt–0.9Ag/0.5K–Al₂O₃, fresh 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ and spent 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ samples.

Considering the 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ sample, the weak peak centered at around 26.15° may be attributed to the presence of a Pt₁Sn₁ intermetallic compound.²⁵ It goes without saying that other Pt/Sn intermetallic compounds may also be present, but the resolution of the X-ray diffractogram does not allow their detection. Peaks centered at around 27.05°, 64.31° and 65.00° have been assigned to SnO₂ moieties.²⁶ Two weak peaks centered at about 43.08° and 54.00° present in the spectra of the 0.3Pt–0.9Ag/0.5K–Al₂O₃ and 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ samples are attributed to a Pt_nAg_m intermetallic compound, where n and m are near 0.5.²¹ Pan et al.²¹ have presented a simulated XRD pattern of an Ag_48.4Pt_51.6 ordered metallic alloy with an HCP phase. They report an enhanced chemical reactivity of the latter intermetallic compound with respect to random Ag/Pt alloys. It should be remarked that characterization of intermetallic structures based solely on weak XRD signals is not complete, and complementary evidence through other analytical techniques like high-resolution transmission electron microscopy is needed.

An interesting general deduction is that in the case of the 0.3Pt/0.5K–Al₂O₃ sample, the Pt element in the metallic or oxide state is almost undetectable in the XRD pattern. The addition of Ag or Sn makes the Pt element detectable (although to a small extent) by forming large clusters of intermetallic compounds.

Referring to the XRD pattern of the spent 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ sample in Fig. 1, it may be inferred that the initial reduction, the PDH process, at the reaction temperature and further cooling under a nitrogen atmosphere to room temperature does not change the phase structure of the composite. In other words, the chemical structure and size of intermetallic compound clusters remain approximately intact.

It should be stated that the use of hexachloroplatinate and tin chloride as raw materials may leave some chlorine atoms on the surface, which may further react with Ag and form AgCl entities. If present, AgCl would have a very low content and not be detectable by the XRD analysis. Its possible existence will be addressed in the XPS characterization section.

H₂-TPR analysis

Fig. 2 shows the H₂-TPR diagram of the fresh the 0.3Pt/0.5K–Al₂O₃, 0.3Pt–0.7Ag/0.5K–Al₂O₃, 0.3Pt–0.7Ag-0.7Sn/0.5K–Al₂O₃, 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃, and 0.3Pt–1.1Ag–0.7Sn/0.5K–Al₂O₃ catalyst samples. The 0.3Pt/0.5K–Al₂O₃ sample exhibits two intense and broad peaks centered at 206 and 293 °C. The first one is attributed to the reduction of platinum oxide and/or platinum oxo-chloride species not in contact with the support, and the second one, in strong interaction with the support.²⁷ Broad peaks of relatively large integrated intensity of platinum oxide species are a feature of highly dispersed Pt particles that provide more exposed Pt atoms prone to transformation into PtO_x/PtO_xCl_y species during the calcination process in air. The TPR profile of the 0.3Pt/0.5K–Al₂O₃ sample is similar to the one reported by Hoang et al.²⁸ for a 0.35Pt/Al₂O₃ sample (0.35 wt% Pt), with the main peaks appearing at 127 and 277 °C.

Fig. 2 H₂-TPR profiles of 0.3Pt/0.5K–Al₂O₃, 0.3Pt–0.7Ag/0.5K–Al₂O₃, 0.3Pt–0.7Ag–0.7Sn/0.5K–Al₂O₃, 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃, and 0.3Pt–1.1Ag–0.7Sn/0.5K–Al₂O₃ catalyst samples.

The diagram of the 0.3Pt–0.7Ag/0.5K–Al₂O₃ sample is somewhat peculiar. The total peak area is significantly lower than for the 0.3Pt/0.5K–Al₂O₃ sample. A peak at 161 °C appears, which is attributed to PtO_x/PtO_xCl_y species reduction. It should be noted that generally PtO_x species reduce under a hydrogen atmosphere at temperatures lower than 200 °C.²⁷ The presence of Ag significantly reduces the first reduction temperature of the Pt oxide species, presumably by diminishing their interaction with the support. Ag₂O supported on Al₂O₃ is known to reduce to metallic Ag at very low temperatures (<50 °C) if in the form of large clusters, i.e., with minimal interaction with the support.²⁹ In addition, Ag₂O species supported by Al₂O₃ have been reported to reduce at near 100 °C.³⁰ Conversely, small Ag₂O clusters interact strongly with the alumina support, and exhibit a reduction temperature in the range of 240–360 °C.²⁹ González Hernández et al.³⁰ report a major peak located at around 340 °C for Ag/Al₂O₃ catalysts and attribute it to Ag₂O clusters. Accordingly, the small peak centered at 343 °C may be attributed to Ag₂O clusters. The remaining peak centered at around 241 °C is attributed to PtO_x/PtO_xCl_y species in strong interaction with the support.

The 0.3Pt–0.7Ag–0.7Sn/0.5K–Al₂O₃ sample is the supported trimetallic compound with the lowest amount of Ag. A shoulder near 100 °C is considered to arise from the reduction of Ag₂O clusters. The broad peak centered at about 160 °C (between 100 and 220 °C) is attributed to the reduction of oxide and oxo-chloride platinum compounds. A shoulder at ca. 245 °C represents PtO_x/PtO_xCl_y species in strong interaction with the support. The major peak located at 323 °C is attributed to discrete SnO₂ clusters on the support.³¹ The shoulder at 340 °C is again related to Ag₂O clusters.³⁰ The small peak centered at 449 °C corresponds to SnO₂ clusters in strong interaction with the support.³¹ Considering the 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ sample, the general trend of the H₂-TPR diagram is similar to that of the 0.3Pt–0.7Ag–0.7Sn/0.5K–Al₂O₃ sample, albeit with a shift to lower temperatures for most of the aforementioned peaks: 173 °C (PtO_x/PtO_xCl_y species), 267 °C (PtO_x/PtO_xCl_y species in strong interaction with the support), 300 °C (discrete SnO₂ clusters), 350 °C (Ag₂O clusters), and 427 °C (SnO₂ clusters in strong interaction with the support). The relative amount of discrete SnO₂ clusters decreases sharply upon increasing the Ag content from 0.7 to 0.9 wt%. Further increasing the Ag content (0.3Pt–1.1Ag–0.7Sn/0.5K–Al₂O₃ sample) results in a lower content of PtO_x/PtO_xCl_y species (178 °C), an insignificant change in the content of PtO_x/PtO_xCl_y species strongly interacting with the support (249 °C), a greater amount of discrete SnO₂ clusters (306 °C), a more Ag₂O clusters (350 °C) and a larger amount of SnO₂ clusters strongly interacting with the support (426 °C).

XPS analysis

XPS profiles of three typical fresh catalyst samples, 0.3Pt/0.5K–Al₂O₃, 0.3Pt–0.9Ag/0.5K–Al₂O₃, and 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃, have been analyzed. Accordingly, it is expected that better scientific insight will be useful in the explanation of the catalytic behavior of the different catalyst samples under consideration.

As a first step, Pt is considered. Based on the XPS spectra obtained, the Pt 4d_5/2 signal was too noisy and weak for all the samples. Accordingly, the Pt 4f_5/2 peak was considered, despite the complexity in its deconvolution due to the overlap with the Al 2p signal originating from the γ-Al₂O₃ support.^32–34Fig. 3 shows the Pt 4f_5/2 spectra. The peaks could be easily deconvoluted into their Pt 4f_5/2 and Al 2p components. All the samples exhibit an Al 2p peak centered at 75.20 eV, in agreement with reported binding energies for alumina samples.^32,35 The consistency of the 2p electron binding energy of the oxide state of aluminum in all the samples is due to the dominance of the excess of the γ-Al₂O₃ support in each of the catalyst composites. Pt in its zero valence state (metallic) has a 4f_5/2 peak in the range of 74.50–74.60 eV.^35,36 The Pt 4f_5/2 peak of the 0.3Pt/0.5K–Al₂O₃ sample appears at 74.72 eV, thus representing a Pt atom with a slightly oxidized character. This state may be due to electronic interactions with the alumina support, as stated before discussing the H₂-TPR diagram of the same sample. By introducing Ag, i.e. the 0.3Pt–0.9Ag/0.5K–Al₂O₃ sample, the Pt 4f_5/2 peak shifts to 74.47 eV, assuming more metallic character. This finding is also in agreement with the H₂-TPR diagram of the 0.3Pt–0.9Ag/0.5K–Al₂O₃ sample and is attributed to less interaction between the Pt clusters and the alumina support. The 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ sample shows a Pt 4f_5/2 peak at 74.70 eV. Upon the introduction of Sn metal into the catalyst composition and Pt–Sn alloy formation, electron transfer from Sn to the more electronegative Pt atoms results in the increase of the Pt 4f_5/2 binding energy level with respect to the 0.3Pt–0.9Ag/0.5K–Al₂O₃ sample.

Fig. 3 XPS spectra of the Pt 4f_5/2 energy level for the 0.3Pt/0.5K–Al₂O₃, 0.3Pt-0.9Ag/0.5K–Al₂O₃, and 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ samples.

Fig. 4 shows the XPS spectra of the Ag 3d_3/2 and Ag 3d_5/2 energy levels of the 0.3Pt–0.9Ag/0.5K–Al₂O₃ and 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ samples. Considering first the 0.3Pt–0.9Ag/0.5K–Al₂O₃ sample, it is observed that the peaks due to Ag 3d_3/2 and Ag 3d_5/2 energy levels are split into two groups: (a) Ag 3d_3/2: 369.5 and 366.98 eV, and (b) Ag 3d_5/2: 376.2 and 373.1 eV. In each group, the lower level represents the metallic status of the Ag element, and the higher level the oxide state.³⁷ The value at 366.98 (rounded to 367.0 eV in Fig. 4) is less than the values for the Ag 3d_3/2 energy levels (368.2 eV) reported for pure metallic Ag.³⁶ Instead, it is near the Ag 3d_3/2 energy level of 367.5 eV of AgBr, where Ag is still in the metallic form, although in polar interaction with Br. The value of 367.0 eV is attributed to strong Ag–Pt interactions,²⁰ resulting in partial electron transfer from metallic Ag to the more electronegative Pt atom. In addition, the satellite peaks indicated as “sp” in Fig. 4 corroborate with those reported in the open literature.³⁶ Based on the XRD results discussed above, Ag₂O constitutes the major oxide component and its percentage (Ag₂O/(Ag⁰ + Ag₂O) × 100) is calculated as 74.16% based on the XPS spectrum.

Fig. 4 XPS spectra of the Ag 3d energy levels for the 0.3Pt–0.9Ag/0.5K–Al₂O₃ and 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ samples. Satellite peaks are indicated as “sp”.

Considering the Ag 3d spectrum of the 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ sample (Fig. 4), it is observed that again Ag takes on both oxide and metallic properties. However, the content of the oxide state decreases to 56.3% with respect to the 0.3Pt–0.7Ag/0.5K–Al₂O₃ sample. Compared to the corresponding peaks of the 0.3Pt–0.9Ag/0.5K–Al₂O₃ sample, the metallic Ag 3d_3/2 and Ag 3d_5/2 energy levels undergo positive shifts of 0.67 and 0.85 eV, respectively. This is attributed to the presence of Sn atoms that reduce the effective electronegativity of the Pt atoms. Referring to Fig. 1, Ag_nPt_m intermetallic compounds may be present alongside metallic Ag⁰ entities for both of the 0.3Pt–0.9Ag/0.5K–Al₂O₃ and 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ samples. In the latter case, the abundance of Sn atoms may increase the metallic character of Ag atoms with respect to the 0.3Pt–0.9Ag/0.5K–Al₂O₃ sample. The electronegativity of Sn atoms is much smaller than that of the Pt atoms.

The adjacency of Sn and Ag atoms may eventually increase the metallic character of Ag atoms in Ag_nPt_m intermetallic compounds by decreasing the net electronegativity difference between Ag and the other heteroatoms. It should be noted that at the nanoscale, Ag and Pt intermetallics may assume a wide and nearly continuous range of compositions, ranging from Pt-poor to Pt-rich structures.²¹ Therefore, the discussed positive observed shift of the metallic Ag 3d_3/2 and Ag 3d_5/2 energy levels may also originate from a different Ag–Pt intermetallic composition in the case of the 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ sample. Fig. 5 shows the XPS spectrum of the Sn 3d energy level of the 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ sample. In the absence of any support interaction, the metallic Sn 3d_5/2 and 3d_3/2 energy levels appear at around 484.3 and 493.0 eV, respectively.³⁸ Based on Fig. 5, these levels appear at 487.1 and 495.8 eV, respectively. The observed positive shifts in the binding energies are the result of Sn oxidation into II and IV states, and are in accordance with reported XPS profiles for Pt–Sn/Al₂O₃ catalysts.^23,38 Although the XRD pattern of the 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ sample showed the presence of the IV oxidation state of Sn, the existence of Sn in the II oxidation state cannot be excluded. Due to the low signal-to-noise ratio of the experimental data, no further deconvolution of the curves was implemented.

Fig. 5 XPS spectrum of the Sn 3d energy levels of the 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ sample.

The XPS survey spectrum did not show any detectable level of Cl atoms. Accordingly, any chloride species, like AgCl, if present, is extremely rare.

Nitrogen adsorption–desorption and NH₃-TPD analysis

Referring to Table 1, it is observed that the specific surface areas and pore volumes of the 0.3Pt/0.5K–Al₂O₃, 0.3Pt–0.9Ag/0.5K–Al₂O₃, and 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ samples are quantitatively similar. The magnitude of the specific surface area of the catalyst samples is close to that of the γ-Al₂O₃ support (197 m² g⁻¹). The relatively low amount of metal loading on the internal surface of the support does not significantly change the specific surface area. Instead, it slightly reduces the pore volume (13%) due to the intercalation of the added material within the pores. The spent catalysts show a 4–8% decrease in specific surface area, probably due to a change in the morphology of the metal components during the reduction step, and not due to sintering effects.

Table 1 Summary of the results obtained from the analysis of nitrogen adsorption/desorption isotherms of the fresh and spent 0.3Pt/0.5 K–Al₂O₃, 0.3Pt–0.9Ag/0.5K–Al₂O₃, and 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ samples

Sample	BET specific surface area (m^2 g^−1)		Pore volume (cm^3 g^−1)
	Fresh	Spent	Fresh	Spent
γ-Al2O3	197	—	0.61	—
0.3Pt/0.5K–Al2O3	192	186	0.53	0.51
0.3Pt–0.9Ag/0.5K–Al2O3	197	180	0.54	0.51
0.3Pt–0.9Ag–0.7Sn/0.5K–Al2O3	191	183	0.52	0.52

---

The NH₃-TPD profiles of the 0.3Pt/0.5K–Al₂O₃, 0.3Pt–0.9Ag/0.5K–Al₂O₃, and 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ samples are shown in Fig. 6. The curves have not been normalized. All the samples show five main peaks in the ranges of 112–113 °C, 175–191 °C, 232–244 °C, 410–419 °C and 675–703 °C, corresponding to very weak, weak, medium, and strong acidic sites, and a dehydroxylation domain, respectively. The latter domain extends well beyond 800 °C, and has been reported to correspond to strongly adsorbed water desorption or dehydroxylation reactions.¹²Table 2 summarizes the deconvolution results based on a Gaussian distribution function. No significant changes in the peak temperatures of each of the acidic sites are observed between the 0.3Pt/0.5K–Al₂O₃, 0.3Pt–0.9Ag/0.5K–Al₂O₃, and 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ samples.

Fig. 6 NH₃-TPD profiles for the 0.3Pt/0.5K–Al₂O₃, 0.3Pt–0.9Ag/0.5K–Al₂O₃, and 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ samples.

Table 2 Acidity-strength distribution details of the 0.3Pt/0.5K–Al₂O₃, 0.3Pt–0.9Ag/0.5K–Al₂O₃, and 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ samples (vw = very weak, w = weak, m = medium, s = strong, DH = dehydroxylation)

Sample	Acidic strength	Temperature (°C)	Area percent	Total area (a.u.)
0.3Pt/0.5K–Al2O3	vw	112.1	0.86	11.03
	w	189.9	20.92
	m	235.5	36.61
	s	410.6	38.93
	DH	703.0	2.86
0.3Pt–0.9Ag/0.5K–Al2O3	vw	112.0	0.65	8.46
	w	175.3	29.14
	m	243.5	26.46
	s	411.5	36.80
	DH	667.7	1.05
0.3Pt–0.9Ag–0.7Sn/0.5K–Al2O3	vw	112.2	3.75	9.38
	w	190.66	17.09
	m	232.32	36.13
	s	418.37	38.16
	DH	675.4	4.86

---

In summary, the acid site strength distribution and the total acidity of the samples are similar. As the main source of acidic sites is the base-modified support, i.e. γ-Al₂O₃ modified with KOH (see the Experimental section), the aforementioned resemblance seems rational.

Electron microscopy analysis

As stated in the previous sections, catalytic evaluation tests were carried out for catalysts in their original shape, i.e. mainly in the shape of the support, to compare their kinetics with available commercial catalysts. Accordingly, it is essential to have a clear picture of the composition profile of each metal element as a function of catalyst particle radius. Fig. 7 shows the cross-section of a catalyst particle (0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃, the optimum sample to be presented in the following sections) and the corresponding element composition profile obtained by EDX line mapping. The values reported are the average values obtained for different line-scans on three radii on the same particle. The radius shown in Fig. 7 is one of the three investigated.

Fig. 7 Cross-section of a catalyst particle (0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃) and the corresponding element composition profile based on EDX analysis.

Fig. 7 clearly shows that the metals Pt, Ag, Sn and K have a radical compositional profile. An egg-shell profile is obvious, where most of the metal content resides in about 5% of the total particle volume. It is noted that K, Pt, Ag and Sn have been added in sequential steps. Nonetheless, the latter elements follow a similar composition profile.

Kinetic performance evaluation

To efficiently evaluate the diversity of the catalytic tests performed, the corresponding experimental results were divided into two general groups: (a) a group of catalysts containing only Pt and Ag on the support, and (b) a group of catalysts containing Pt, Ag and Sn. Considering group (a), we aim to investigate the influence of the presence of Ag as the second metal and its content (from 0.0 to 1.1 wt%) on the PDH reaction. Considering group (b), the effect of adding a constant amount of Sn (0.7 wt%) to the catalysts of group (a) on the PDH reaction is studied.

Fig. 8a–c show the conversion, propylene selectivity and propylene yield of the 0.3Pt–xAg/0.5K–Al₂O₃ group of catalysts, where x assumes values of 0.0, 0.1, 0.3, 0.5, 0.7, 0.9 and 1.1. Referring to Fig. 8a, it is observed that the final conversion at 200 min significantly increases with increasing Ag content from 0.0 to 0.7 wt%, and then decreases slightly up to an Ag content of 1.1 wt%. A quantitative illustration of the deactivation rate is better shown in Fig. 9a. It is observed that the deactivation constant k_d is at a maximum value for the 0.0 Ag wt% catalysts (5.561 × 10⁻² min⁻¹) and decreases to a minimum value with increasing Ag content up to 0.5 wt% (5.125 × 10⁻² min⁻¹). Increasing the Ag content from 0.5 to 1.1 wt% results in a slight increase up to 5.230 × 10⁻² min⁻¹. Referring to Fig. 8b, it is observed that after 200 min, the propylene selectivity increases from a minimum of 69% for 0.0 Ag wt% to 72% for 0.7 Ag wt%. A further increase in the Ag content to 0.9 wt% results in an abrupt increase in the propylene selectivity to 81%. A further increase in the Ag content to 1.1 wt% results in a slight decrease in the selectivity to 79%. Considering Fig. 8c, it is observed that the propylene yield steadily increases from an Ag content of 0.0 to 0.9 wt% from 19 to 28%, and moderately increases to 29% for an Ag content of 1.1 wt%. Fig. 9b shows the trend described above.

Fig. 8 Conversion (a), propylene selectivity (b) and propylene yield (c) of the 0.3Pt–xAg/0.5K–Al₂O₃ group of catalysts, where x assumes values of 0.0, 0.1, 0.3, 0.5, 0.7, 0.9 and 1.1 wt%.

Fig. 9 a) Deactivation constant k_d as a function of Ag wt% with and without 0.7 wt% Sn, and b) propylene yield as a function of Ag wt% with and without 0.7 wt% Sn.

In summary, it is observed that in the absence of Sn, Ag acts as a strong promoter for 0.3Pt/0.5K–Al₂O₃ base catalysts. Although up to an Ag content of 1.1 wt%, a steady increase in yield is observed, the minimum deactivation rate occurs at an Ag content of 0.5 wt%. To remediate, the effect of the addition of a constant amount of 0.7 wt% Sn to all the catalysts of group “a” is further considered.

Fig. 10 shows the effect of Ag concentration on conversion, propylene selectivity and propylene yield for the 0.3Pt–xAg–0.7Sn/0.5K–Al₂O₃ group of catalysts, where x assumes values of 0.0, 0.1, 0.3, 0.5, 0.7, 0.9 and 1.1. The addition of 0.7 wt% Sn results in a substantial increase in the final conversion at 200 min from 27 to 37% for the 0.3Pt/0.5K–Al₂O₃ base catalyst in the absence of Ag addition. The addition of 0.1 wt% Ag abruptly decreases the conversion down to 30%, and further increase in Ag content to 0.7 wt% results in a steady increase of conversion up to 39%. Further increase in Ag content to 1.1 wt% results in a steady decrease of conversion down to 37%. Referring again to Fig. 9a, it is observed that the lowest deactivation rate for the 0.3Pt–xAg–0.7Sn/0.5K–Al₂O₃ catalyst series is for an Ag content of 0.9 wt% (5.076 × 10⁻² min⁻¹). Considering propylene selectivity, only Sn addition to the base 0.3Pt/0.5K–Al₂O₃ catalyst results in a significant increase from 69 to 77%. The addition of 0.1 wt% Ag abruptly reduces selectivity down to 72%. Further increase in Ag content to 0.5 wt% results in an increase in the selectivity up to 77%. The selectivity reduces sharply to 69% upon increasing the Ag content to 0.7 wt%. A further increase in Ag content to 1.1 wt% results in a steady increase in the propylene selectivity up to 81%. Considering Fig. 10c, it is observed that the propylene yield assumes a relatively high value of 29% for the 0.3Pt–0.0Ag–0.7Sn/0.5K–Al₂O₃ catalyst. The addition of 0.1 wt% Ag causes a sharp decrease in the yield to 21%. However, the yield steadily increases with a further increase in Ag content and reaches a plateau of 30% at an Ag content of 1.1 wt%. The trend of propylene yield as a function of Ag content in the presence of Sn is shown in Fig. 9b.

Fig. 10 Conversion (a), propylene selectivity (b) and propylene yield (c) of the 0.3Pt–xAg–0.7Sn/0.5K–Al₂O₃ group of catalysts, where x assumes values of 0.0, 0.1, 0.3, 0.5, 0.7, 0.9 and 1.1. The corresponding data for the DeH-16 catalysts at 200 min on stream is shown on the left with arrows.

It may be stated that in the presence of Sn, Ag acts as a strong promoter for the 0.3Pt–xAg–0.7Sn/0.5K–Al₂O₃ group of catalysts only when the Ag content is equal or greater than 0.9 wt%. As far as the propylene yield and deactivation rate are concerned, the best performance among the 0.3Pt–xAg–0.7Sn/0.5K–Al₂O₃ group of catalysts is exhibited by the 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ catalyst. Referring to Fig. 9, it may also be stated that the 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ catalyst exhibits the best performance among all the 0.3Pt–xAg/0.5K–Al₂O₃ and 0.3Pt–xAg–0.7Sn/0.5K–Al₂O₃ groups of catalysts. Remarkably, a DeH-16 commercial catalyst resulted in a conversion, propylene selectivity and propylene yield of 41, 80 and 32%, respectively, under the same operating conditions, comparable to the performance of our optimum sample.

Before trying to explain the observed catalytic behavior as a function of Ag content in the presence and absence of Sn, it would be informative to consider briefly the TPO analysis of the “spent” 0.3Pt/0.5K–Al₂O₃, 0.3Pt–0.7Ag/0.5K–Al₂O₃, and 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ samples (200 min time on stream).

TPO and Raman analysis of spent catalysts

Fig. 11 illustrates the TPO profiles of the 0.3Pt/0.5K–Al₂O₃, 0.3Pt–0.9Ag/0.5K–Al₂O₃, and 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ samples after 200 min on stream. The 0.3Pt/0.5K–Al₂O₃ sample exhibits two peak maxima at 389 and 500 °C. The first peak is attributed to the oxidation of coke formed on the active Pt metallic sites, and the second one to the oxidation of carbon formed on the support.^23,31 The addition of Ag (0.3Pt–0.9Ag/0.5K–Al₂O₃ sample) practically eliminates the second peak and shifts the first one to the lower value of 361 °C. In addition, the integrated intensity due to the first peak significantly decreases. The optimum sample, 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃, exhibits again one dominant peak at low temperatures (373 °C), but with an integrated intensity substantially smaller than that of the 0.3Pt–0.9Ag/0.5K–Al₂O₃ sample. Accordingly, it may be stated that the optimum trimetallic sample produces the lowest amount of carbon and mainly on the metal sites. The structural properties of deposited coke on the 0.3Pt/0.5K–Al₂O₃, 0.3Pt–0.9Ag/0.5K–Al₂O₃, and 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ samples after 200 min on stream may be assessed by Raman analysis (Fig. 12). All the catalysts exhibit peaks at around 1330 and 1609 cm⁻¹. The peaks at 1330 cm⁻¹ (D band) and 1609 cm⁻¹ (G band) are attributed to the aromatic C–H and C=C bond vibrations, respectively, and represent typical bonds in graphite-like carbon materials.¹⁰Fig. 12 also shows the deconvoluted spectra resulting in individual D1, D3, D4 and G bands. The I_D/I_G intensity ratios of the spent 0.3Pt/0.5K–Al₂O₃, 0.3Pt–0.9Ag/0.5K–Al₂O₃, and 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ samples are calculated to be 0.77, 0.7, and 0.7, respectively. An intensity ratio of 0.8 has previously been reported for graphite-like carbon.^39–41 A 9% decrease in the I_D/I_G intensity ratio upon addition of Ag and Sn metals shows a decrease in the crystallinity of the deposited graphite-like carbon. In addition, the total integrated intensity in the range of 1000–2000 cm⁻¹ decreases sharply upon the addition of Ag and Sn metals, and this is in accordance with the TPO results discussed previously.

Fig. 11 TPO profiles of the 0.3Pt/0.5K–Al₂O₃, 0.3Pt–0.9Ag/0.5K–Al₂O₃, and 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ samples after 200 min on stream.

Fig. 12 Raman spectra of the spent 0.3Pt/0.5K–Al₂O₃, 0.3Pt–0.9Ag/0.5K–Al₂O₃, and 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ samples after 200 min time on stream. (D1, D3, D4 and G appear at 1337, 1536, 1166 and 1608 cm⁻¹, respectively.)

Based on the detailed discussions regarding fresh and spent catalyst characterization and catalytic performance evaluation, a general picture may be developed. Less coke is produced on the 0.3Pt–0.9Ag/0.5K–Al₂O₃ catalyst due to the presence of Pt_nAg_m intermetallics. The 0.3Pt–0.9Ag–0.7Sn/0.5 K–Al₂O₃ is subject to the least coking due to the presence of two kinds of bimetallics, Pt_nAg_m and Pt_xSn_y. It is argued that there exists a clear relationship between coking depression and selectivity enhancement toward propylene. Part of the catalyst deactivation is due to the support. Considering the optimum sample, 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃, based on XRD analysis, it is most probable that extra Ag and Sn atoms not forming alloys with Pt atoms cover the support surface, eventually reducing its acidic character. Fig. 13 is a schematic drawing that emphasizes the high tendency of monometallic Pt clusters to produce coke (zone A), while the Pt–Ag (zone B) and Pt–Sn (zone C) alloy clusters, especially the intermetallic ones, have a low tendency for deep hydrogenolysis. The hydrogenolysis reaction is accompanied by methane, ethane and ethylene production along with coke formation. Because of that, selectivity is higher near the bi and trimetallic alloy clusters.

Fig. 13 Schematic drawing showing the effect of Ag and Sn promoters in the optimum 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ trimetallic catalyst on coke formation and selectivity versus propylene production (black = C (coke), blue = Pt, red = Ag, cyan = Sn, gray = C, white = H). See the text for descriptions of the A–C zones. The proportions of Pt, Ag and Sn atoms are not exact.

It should be noted that the stoichiometric formulas of the catalysts refer to nominal values. It is essential to report these nominal values because they represent the stoichiometry used in the synthesis of the catalysts. The reader can therefore reproduce the synthesis method described in the Experimental section. XRF analysis of the optimum sample revealed 0.27, 0.52 and 0.66 values for the Pt, Sn and Ag contents of the optimum sample, respectively.

Finally, it would be interesting to compare the catalytic performance of the optimum sample with reported data. Gong and Zhao⁴² report a conversion in the range of 21–30% and selectivity in the range of 35–85% for a PtSn/Al₂O₃ catalyst for GSHV = 20 000 cm³ g_cat⁻¹ h⁻¹ for a feed composition of 26 vol% C₃H₈ (other gases not specified) at a reaction temperature of 600 °C. Sun et al.⁴³ report a conversion in the range of 33–43% and selectivity close to 98% for a PtSn/Al₂O₃ catalyst for GSHV = 15 000 cm³ g_cat⁻¹ h⁻¹ for a feed composition of C₃H₈ : H₂ : N₂ = 3 : 1 : 21 at a reaction temperature of 540 °C. Due to different operating conditions applied and different catalyst reduction methods, a direct comparison is difficult. In addition, the references cited used crushed particles and not industrial-size spherical pellets.

Conclusions

The outcomes of the present study fill a gap in the open literature regarding the use of a Ag promoter in Pt/γ-Al₂O₃ and Pt–Sn/γ-Al₂O₃ catalysts. More specifically, Pt–Ag–Sn trimetallic catalysts for the PDH reaction are considered for the first time. It has been shown that for a constant 0.3 wt% Pt content, the propylene yield steadily increases with Ag content up to 1.1 wt%. However, the best catalyst stability against deactivation is obtained at an Ag content of 0.5 wt%. The same increasing trend in the propylene yield with increasing Ag content up to 1.1 wt% is observed for the trimetallic case (constant Sn content of 0.7 wt%). In the latter case, however, the lowest deactivation is observed for an Ag content of 0.9 wt%. The catalytic performance of the optimum catalyst, i.e. the 0.3Pt–0.9Ag–0.7Sn/0.5K–Al₂O₃ sample, is comparable with that of DeH-16, a commercial dehydrogenation catalyst. It is remarkable that the DeH-16 commercial catalyst resulted in a conversion, propylene selectivity and propylene yield of 41, 80 and 32%, respectively, under the same operating conditions, comparable to the performance of our optimum sample. The trimetallic catalyst may still be improved by changing the Sn content, and this is the subject of our future studies.

Author contributions

G. K.: investigation, conceptualization, methodology, visualization, writing – original draft. C. F.: investigation, conceptualization, methodology, visualization, project administration, supervision, validation, writing – original draft, writing – review and editing. M. M.: investigation, conceptualization, methodology, visualization, project administration, supervision, validation, writing – review and editing. R. D.: investigation, methodology, writing – review and editing.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The authors of this manuscript state that all the data associated with this article are included in the manuscript text.

References

1. Z.-P. Hu, D. Yang, Z. Wang and Z.-Y. Yuan, Chin. J. Catal., 2019, 40, 1233–1254.
2. S. Gómez-Quero, T. Tsoufis, P. Rudolf, M. Makkee, F. Kapteijn and G. Rothenberg, Catal. Sci. Technol., 2013, 3, 962–971.
3. C. Zuo and Q. Su, Molecules, 2023, 28, 3594.
4. G. Festa, A. Serrano-Lotina, E. Meloni, R. Portela, C. Ruocco, M. Martino and V. Palma, Ind. Eng. Chem. Res., 2024, 63, 16269–16284.
5. G. Festa, P. Contaldo, M. Martino, E. Meloni and V. Palma, Ind. Eng. Chem. Res., 2023, 62, 16622–16637.
6. M. Martino, E. Meloni, G. Festa and V. Palma, Catalysts, 2021, 11, 1070.
7. P. Praserthdam, T. Mongkhonsi, S. Kunatippapong, B. Jaikaew and N. Lim, in Studies in Surface Science and Catalysis, Elsevier, 1997, vol. 111, pp. 153–158.
8. O. A. Bariås, A. Holmen and E. A. Blekkan, Catal. Today, 1995, 24, 361–364.
9. Y. Wang, Y. Wang, S. Wang, X. Guo, S.-M. Zhang, W.-P. Huang and S. Wu, Catal. Lett., 2009, 132, 472–479.
10. G.-Q. Ren, G.-X. Pei, Y.-J. Ren, K.-P. Liu, Z.-Q. Chen, J.-Y. Yang, Y. Su, X.-Y. Liu, W.-Z. Li and T. Zhang, J. Catal., 2018, 366, 115–126.
11. L. G. Cesar, C. Yang, Z. Lu, Y. Ren, G. Zhang and J. T. Miller, ACS Catal., 2019, 9, 5231–5244.
12. H. Lee, W.-I. Kim, K.-D. Jung and H. L. Koh, Korean J. Chem. Eng., 2017, 34, 1337–1345.
13. G. Sun, Z.-J. Zhao, R. Mu, S. Zha, L. Li, S. Chen, K. Zang, J. Luo, Z. Li and S. C. Purdy, Nat. Commun., 2018, 9, 4454.
14. S. Sun, G. Sun, C. Pei, Z.-J. Zhao and J. Gong, J. Phys. Chem. C, 2021, 125, 18708–18716.
15. L.-L. Shen, K. Xia, W.-Z. Lang, L.-F. Chu, X. Yan and Y.-J. Guo, Chem. Eng. J., 2017, 324, 336–346.
16. L.-L. Long, K. Xia, W.-Z. Lang, L.-L. Shen, Q. Yang, X. Yan and Y.-J. Guo, J. Ind. Eng. Chem., 2017, 51, 271–280.
17. Z. Wu, B. C. Bukowski, Z. Li, C. Milligan, L. Zhou, T. Ma, Y. Wu, Y. Ren, F. H. Ribeiro and W. N. Delgass, J. Am. Chem. Soc., 2018, 140, 14870–14877.
18. W. Cai, R. Mu, S. Zha, G. Sun, S. Chen, Z.-J. Zhao, H. Li, H. Tian, Y. Tang and F. Tao, Sci. Adv., 2018, 4, eaar5418.
19. T. Wang, F. Jiang, G. Liu, L. Zeng, Z. j. Zhao and J. Gong, AIChE J., 2016, 62, 4365–4376.
20. M. T. Schaal, M. P. Hyman, M. Rangan, S. Ma, C. T. Williams, J. R. Monnier and J. W. Medlin, Surf. Sci., 2009, 603, 690–696.
21. Y.-T. Pan, Y. Yan, Y.-T. Shao, J.-M. Zuo and H. Yang, Nano Lett., 2016, 16, 6599–6603.
22. G. Kasaeian, C. Falamaki and M. Mozaffariana, Iran. J. Sci., 2025, 1–16.
23. X. Gao, W. Xu, X. Li, J. Cen, Y. Xu, L. Lin and S. Yao, Chem. Eng. J., 2022, 443, 136393.
24. J. Fowsiya and G. Madhumitha, J. Cluster Sci., 2019, 30, 1243–1252.
25. Z. He, J. Yang and L. Liu, JACS Au, 2024, 4, 4084–4109.
26. M. Akram, A. T. Saleh, W. A. W. Ibrahim, A. S. Awan and R. Hussain, Ceram. Int., 2016, 42, 8613–8619.
27. Y. Zhang, C. Y. Zhao, H. Liang and Y. Liu, Catal. Lett., 2009, 127, 339–347.
28. D. Hoang, S.-F. Farrage, J. Radnik, M.-M. Pohl, M. Schneider, H. Lieske and A. Martin, Appl. Catal., A, 2007, 333, 67–77.
29. K. Bethke and H. Kung, J. Catal., 1997, 172, 93–102.
30. N. N. González Hernández, J. L. Contreras, M. Pinto, B. Zeifert, J. L. Flores Moreno, G. A. Fuentes, M. E. Hernández-Terán, T. Vázquez, J. Salmones and J. M. Jurado, Catalysts, 2020, 10, 1212.
31. S. A. Bocanegra, S. R. de Miguel, I. Borbath, J. L. Margitfalvi and O. A. Scelza, J. Mol. Catal. A: Chem., 2009, 301, 52–60.
32. P. Peng, J. Li, S. Mo, Q. Zhang, T. Shen and Q. Xie, Processes, 2021, 9, 230.
33. Y. S. Choi, K. Oh, K.-D. Jung, W.-I. Kim and H. L. Koh, Catalysts, 2020, 10, 898.
34. A. Ivanova, E. Slavinskaya, R. Gulyaev, V. Zaikovskii, O. Stonkus, I. Danilova, L. Plyasova, I. Polukhina and A. Boronin, Appl. Catal., B, 2010, 97, 57–71.
35. T. Tuguhiro, N. Kataoka, H. Tanaka, K. Kinoshita and S. Kishida, Procedia Eng., 2017, 216, 175–181.
36. N. J. Firet, M. A. Blommaert, T. Burdyny, A. Venugopal, D. Bohra, A. Longo and W. A. Smith, J. Mater. Chem. A, 2019, 7, 2597–2607.
37. P. K. Parashar and V. K. Komarala, Sci. Rep., 2017, 7, 12520.
38. K. Balakrishnan and J. Schwank, J. Catal., 1991, 127, 287–306.
39. L. Shi, G. M. Deng, W. C. Li, S. Miao, Q. N. Wang, W. P. Zhang and A. H. Lu, Angew. Chem., Int. Ed., 2015, 54, 13994–13998.
40. F. Tuinstra and J. L. Koenig, J. Chem. Phys., 1970, 53, 1126–1130.
41. Z. Zeng and K. Natesan, Chem. Mater., 2003, 15, 872–878.
42. N. Gong and Z. Zhao, Mol. Catal., 2019, 477, 110543.
43. C. Sun, J. Luo, M. Cao, P. Zheng, G. Li, J. Bu, Z. Cao, S. Chen and X. Xie, J. Energy Chem., 2018, 27, 311–318.

---