High-purity hydrogen production from dehydrogenation of methylcyclohexane catalyzed by zeolite-encapsulated subnanometer platinum-iron clusters

He, Zhe; Li, Kailang; Chen, Tianxiang; Feng, Yunchao; Villalobos-Portillo, Eduardo; Marini, Carlo; Lo, Tsz Woon Benedict; Yang, Fuyuan; Zhang, Liang; Liu, Lichen

doi:10.1038/s41467-024-55370-z

Download PDF

Article
Open access
Published: 02 January 2025

High-purity hydrogen production from dehydrogenation of methylcyclohexane catalyzed by zeolite-encapsulated subnanometer platinum-iron clusters

Nature Communications volume 16, Article number: 92 (2025) Cite this article

14k Accesses
43 Citations
Metrics details

Subjects

Abstract

Liquid organic hydrogen carriers (LOHCs) are considered promising carriers for large-scale H₂ storage and transportation, among which the toluene-methylcyclohexane cycle has attracted great attention from industry and academia because of the low cost and its compatibility with the current infrastructure facility for the transportation of chemicals. The large-scale deployment of the H₂ storage/transportation plants based on the toluene-methylcyclohexane cycle relies on the use of high-performance catalysts, especially for the H₂ release process through the dehydrogenation of methylcyclohexane. In this work, we have developed a highly efficient catalyst for MCH dehydrogenation reaction by incorporating subnanometer PtFe clusters with precisely controlled composition and location within a rigid zeolite matrix. The resultant zeolite-encapsulated PtFe clusters exhibit the up-to-date highest reaction rate for dehydrogenation of methylcyclohexane to toluene, very high chemoselectivity to toluene (enabling the production of H₂ with purity >99.9%), remarkably high stability (>2000 h) and regenerability over consecutive reaction-regeneration cycles.

Methyl formate as a hydrogen energy carrier

Article Open access 22 June 2023

Catalytic acceptorless complete dehydrogenation of cycloalkanes

Article Open access 09 January 2025

Achieving gigawatt-scale green hydrogen production and seasonal storage at industrial locations across the U.S

Article Open access 19 October 2024

Introduction

With the explosive production of green hydrogen from water electrolyzers based on renewable energy resources, the large-scale storage and transportation of hydrogen are vital for building a sustainable society based on hydrogen energy^1,2. Among the alternative strategies, H₂ storage based on liquid organic hydrogen carriers (LOHC) is considered a promising route for large-scale implementation due to its compatibility with current industry infrastructure^3,4,5,6. To date, several LOHC-based hydrogen storage systems have been proposed, among which the hydrogenation-dehydrogenation cycle between toluene (TOL) and methylcyclohexane (MCH) has attracted tremendous attention from academia and industry^7,8. As shown in Fig. 1a, the combination of water electrolysis, synthesis of TOL from fossil resources or hydrogenation of captured CO₂, and the H₂ storage process based on the MCH-TOL cycle can establish a sustainable production chain in the generation and transportation of green H₂ and aromatic resources, which strongly relies on the development of efficient catalysts in each step.

Fig. 1: Solid catalysts for H2 storage and transportation based on toluene-methylcyclohexane cycle. — **Fig. 1: Solid catalysts for H₂ storage and transportation based on toluene-methylcyclohexane cycle.**

From the mechanistic point of view, the dehydrogenation of MCH to TOL could be a critical step in the H₂ storage system based on the MCH-TOL cycle, because the dehydrogenation of MCH is an endothermic process and relies on the activation of sp³-type C-H bonds⁹. From a practical point of view, the solid catalysts need to deal with concentrated MCH feeds to achieve high efficiency in H₂ production. Besides, considering the inevitable deactivation of supported metal catalysts under dehydrogenation reaction conditions due to coke formation, the catalyst must show robust recyclability in consecutive reaction-regeneration cycles. Moreover, the purity of H₂ produced from the dehydrogenation of LOHC is an essential criterion for evaluating the catalyst’s potential for storing H₂ for fuel cells, thus requiring the deployment of a highly selective catalyst for the dehydrogenation process¹⁰. As illustrated in Fig. 1b, although numerous strategies have been proposed for improving the performances of supported metal catalysts for MCH dehydrogenation, such as optimization of metal-support synergy, and modifying the chemical composition and domain size of the metal active sites^{11,12,13,14,15}. For instance, by selectively poisoning the Pt nanoparticles supported on Al₂O₃ by sulfur, high activity, and selectivity have been achieved in the MCH dehydrogenation reaction and the S-modified Pt/Al₂O₃ catalysts are now being used in the commercial process^16,17. To further decrease the cost of the supported Pt catalysts and reduce the reactor size, it is of interest to develop efficient Pt catalysts with high specific activities in H₂ production from MCH dehydrogenation reaction.

In this work, we have developed a highly efficient supported Pt catalyst for MCH dehydrogenation reaction by incorporating subnanometer bimetallic PtFe clusters with precisely controlled composition and location within a zeolite matrix (Fig. 1c). The rigid zeolite framework can provide effective confinement on the subnanometer PtFe clusters to suppress the sintering of the active sites during the MCH dehydrogenation reaction¹⁸. The resultant zeolite-encapsulated PtFe clusters exhibit the up-to-date highest reaction rate for MCH dehydrogenation reaction, very high chemoselectivity to toluene (enabling the production of H₂ with purity >99.9%), remarkably high stability (>2000 h) and regenerability over consecutive reaction-regeneration cycles.

Results

Synthesis of Pt-based catalysts

Pure-silica MFI-type zeolite with three-dimensional pore structures is chosen as the host for subnanometer metal species because the pore dimensions of the 10-member ring (10MR) channels well match the molecule size of TOL and MCH (see illustration in Fig. 1c). We anticipate that, if the active Pt sites can be well stabilized in the 10MR channels of MFI zeolite, the MCH can diffuse along the zeolite channels to access the Pt sites for catalytic transformation into toluene, but the mobility of the active Pt sites will be constrained by the rigid zeolite channels, resulting in a highly stable catalyst^19,20. Following this hypothesis, we have prepared a series of Pt@MFI and bimetallic PtM_4.7@MFI (M = Sn, Zn, Ga, In and Fe) materials one-pot synthesis with an M/Pt molar ratio of ~4.7 (see illustrative description of the synthesis process in Supplementary Fig. 1), which allows the simultaneous encapsulation of subnanometer Pt and M species into the zeolite crystallites during the crystallization process without affecting the formation of MFI zeolite crystallites (see field emission scanning electron microscopy images and XRD patterns in Supplementary Figs. 2–9). The M elements are involved in tuning the electronic properties of Pt species and then modulating the catalytic properties^21,22.

As shown in Fig. 1 and Supplementary Figs. 10–16, subnanometer Pt clusters with an average particle size of ~0.6 nm are formed within the MFI zeolite crystallites in the monometallic Pt@MFI sample, which remain stable after calcination in air and reduction by H₂ at 600 °C. Furthermore, the encapsulations of subnanometer PtM clusters into the MFI zeolite crystallites are also validated with the PtM_4.7@MFI materials, indicating that introducing M elements does not affect the generation and size distributions of Pt clusters.

We have employed the high-resolution scanning transmission electron microscopy (STEM) imaging technique to figure out the exact location of the Pt clusters in the microporous zeolite matrix. As shown in Fig. 2 and Supplementary Figs. 17–30, by correlating the paired high-angle annular dark-field (HAADF-STEM) image and the integrated differential phase contrast (iDPC-STEM) image, the Pt clusters are determined to be predominantly in the sinusoidal 10MR channels of MFI zeolite structure, regardless of the type of the M elements²³. The above characterization results infer that the PtM_4.7@MFI materials can be employed as model systems to study the influences of the chemical composition of subnanometer Pt-based catalysts for MCH dehydrogenation, which may provide new insights into the structure-reactivity relationship in the subnanometer regime.

**Fig. 2: Determination of the location of subnanometer Pt clusters by electron microscopy characterization.**

Evaluation of the Pt-based catalysts for MCH dehydrogenation

The catalytic performances of the Pt-based catalysts have been tested for MCH dehydrogenation reaction in a fix-bed reactor at ambient pressure. As shown in Fig. 3a, b, the monometallic Pt@MFI sample shows nearly complete conversion of MCH at 350 °C at WHSV of 2.2 h^-1 when using mixture of MCH and N₂ (MCH:N₂ = 1:1) as the reaction feed and the Pt@MFI catalyst remains stable for >60 h and gives 97–99% selectivity to TOL over the testing period, suggesting that subnanometer Pt clusters are highly active for MCH dehydrogenation reaction, which could be associated with the advantages of the under-coordinated Pt sites in Pt clusters for activation of saturated C-H bonds in MCH²⁴.

**Fig. 3: Catalytic results of Pt-zeolite materials for MCH dehydrogenation reaction.**

As suggested by Fig. 3a, b, the addition of M species into the Pt@MFI material has remarkable impacts on the catalytic properties. In the cases of PtZn_4.7@MFI, PtGa_4.7@MFI, and PtIn_4.7@MFI, they exhibit significantly declined activities for MCH dehydrogenation. Morphological and spectroscopy characterization of the spent catalysts infer that the low initial reaction rates and rapid catalyst deactivation of PtZn_4.7@MFI, PtGa_4.7@MFI, and PtIn_4.7@MFI catalysts could be related to the modification of the electronic properties of Pt clusters by the co-catalyst metal species, resulting in low capability for C-H activation and elevated coke deposition (Supplementary Figs. 31–35). In contrast, the PtSn_4.7@MFI and PtFe@MFI show similar activity as the Pt@MFI catalyst over the course of ~20 h. More importantly, under steady reaction conditions, the PtSn_4.7@MFI and PtFe_4.7@MFI show improved selectivity to toluene (~99%), indicating that introducing Sn and Fe can suppress the undesired conversion of toluene to benzene. The short induction period observed with the PtFe_4.7@MFI catalyst could be associated with the covering of the residual acid sites by the coke species as indicated by the NH₃ temperature-programmed desorption (NH₃-TPD) profiles (Supplementary Figs. 36). The long-term catalytic tests show that the PtSn_4.7@MFI catalyst suffers deactivation after ~20 h on stream (Fig. 3c, d). In comparison, the PtFe_4.7@MFI catalyst maintains a high conversion of MCH (>90%) and very high selectivity (~99%) for >85 h of time on stream, indicating the potential of PtFe_4.7@MFI as an efficient, stable, and highly selective catalyst for MCH dehydrogenation reaction. The accelerated deactivation of PtSn_4.7@MFI catalyst after reaction for ~25 h could be caused by coke deposition on the MFI zeolite crystallites, leading to a rapid loss of available Pt sites to the reactants.

It should be noted that, besides the co-catalyst metal, K⁺ is also a critical promotor in the Pt-zeolite catalysts for stabilizing the subnanometer Pt clusters in MFI zeolite channels. In the absence of K⁺ promotor, Pt nanoparticles are formed in the K-free PtFe_4.7@MFI catalyst and exhibit very low activity for the MCH dehydrogenation reaction (Supplementary Figs. 37–38).

Structural characterizations of the PtFe@MFI catalysts

Based on the above results, we have optimized the bimetallic PtFe_x@MFI catalysts by varying the Fe loading while fixing the Pt loading. Electron microscopy characterizations of a series of PtFe@MFI catalysts show that adding an excess amount of Fe species as a promotor will cause the sintering of Pt species into large nanoparticles in the PtFe_7.9@MFI catalyst (Fig. 4a), which can cause a dramatic decrease in activity for MCH dehydrogenation reaction. For the PtFe_x@MFI (x = 0.6–4.7) samples with Fe loadings of <0.6 wt%, the morphology of the PtFe_x@MFI samples, size distributions, and dispersion of the Pt species and their locations within the MFI zeolite structure are quite similar (Supplementary Figs. 39–53, Fig. 4b–n and Supplementary Table 4), indicating that introducing an appropriate amount of Fe species will not affect the generation and stabilization of subnanometer Pt clusters in MFI zeolite.

Fig. 4: Electron microscopy characterization of PtFex@MFI catalysts with different compositions. — **Fig. 4: Electron microscopy characterization of PtFe_x@MFI catalysts with different compositions.**

To decipher the atomic-level structural features of the bimetallic PtFe species in different samples, we have employed in situ X-ray absorption spectroscopy to study the chemical states and coordination environment of Pt and Fe species in the working PtFe_x@MFI catalysts. As shown in Fig. 5a, the Pt L₃-edge X-ray absorption near-edge structure (XANES) spectra show the presence of metallic Pt in all the measured Pt-zeolite materials, regardless of their chemical composition, being consistent with the XPS results (Supplementary Fig. 54). The fitting results of the Pt L₃-edge extended X-ray absorption fine structure (EXAFS) spectra (Fig. 5c, Supplementary Fig. 55, and Table 1) indicate the formation of subnanometer Pt clusters in the monometallic Pt@MFI and the PtFe_x@MFI (x = 0.6–4.7) samples while the formation of Pt nanoparticles (>2 nm) is validated with the PtFe_7.9@MFI sample²⁵. It should be noted that when the Fe loading is <0.6 wt%, only minor differences in the coordination number of Pt-Pt bonding are observed with the Pt@MFI and PtFe_x@MFI samples, which is consistent with the electron microscopy characterization results. The low contribution of Pt-Fe bonding in the Pt L₃-edge EXAFS spectra could be caused by the relatively low percentage of Fe species in the bimetallic PtFe clusters²⁶.

**Fig. 5: Spectroscopy characterizations of the Pt-zeolite materials.**

Table 1 Fit results for the Pt L₃-edge and Fe K-edge EXAFS spectra of PtFe-zeolite materials

Full size table

The comparison of the Fe K-edge XANES spectra (Fig. 5b) and those of the reference samples (Fe, FeO, Fe₂O₃ and Fe₃O₄) indicate the formation of reduced FeO_x species in the PtFe@MFI samples after the in situ reduction treatment by H₂ at 600 °C. Interestingly, the K-edge XANES spectra of the PtFe@MFI samples don’t resemble any of the reference samples, inferring the presence of a mixture of different types of Fe species, which is reasonable for Fe-containing zeolite materials because of the low reducibility of Fe species encapsulated in zeolites^27,28. Only the first-shell Fe-O bonding is observed in the Fe K-edge EXAFS spectra (Fig. 5d, Supplementary Fig. 55, and Table 1), regardless of the Fe loading in the PtFe_x@MFI samples, implying the presence of highly dispersed FeO_x species in the form of atomically dispersed Fe species or partially reduced FeO_x clusters in the zeolite matrices, as confirmed by the EDS mapping results (Supplementary Fig. 56)²⁹. The chemical states and coordination environment of Fe species are further studied by quasi in situ Mössbauer spectroscopy³⁰. As summarized in Fig. 5e–h and Supplementary Table 5, regardless of the composition of the PtFe_x@MFI catalysts, the iron species mostly exist as iron oxide species in a combination of tetrahedrally coordinated FeO_x species (FeO_x-Tetra) and hexagonally coordinated FeO_x species (FeO_x-Hexa), which is in line with the measurement results of the magnetic properties (Supplementary Fig. 57). Besides the FeO_x species, we have also observed the presence of partially reduced Fe species in the Mössbauer spectra, which can be associated with the FeO_x (x < 1) clusters in the MFI zeolite channels and the reduced Fe species interacting with Pt clusters (i.e., the FePt bimetallic species). The abundances of the partially reduced Fe species in the PtFe_4.7@MFI and PtFe_7.9@MFI samples are considerably higher than that in the PtFe_2.4@MFI, which is consistent with the chemical composition analysis of the bimetallic clusters/nanoparticles by electron microscopy (Supplementary Fig. 58–59 and Supplementary Tables 6-7) and the UV-vis spectroscopy measurements (Supplementary Fig. 60–61)^27,31. It should be noted that the absence of the FeC_x species in the bimetallic PtFe@MFI catalysts is confirmed by the Mössbauer spectra of the spent catalysts (Supplementary Fig. 62a). The incorporation of Fe into Pt clusters will cause the modification of their electronic properties, as inferred by the IR spectra using CO as the probe molecules (Supplementary Fig. 63), which can influence the adsorption/interaction between the Pt active sites and the reactant/product.

We have followed the reduction process of the pristine PtFe_4.7@MFI sample obtained from one-pot synthesis by in situ XAS to gain further insights into the formation mechanism of subnanometer PtFe clusters. As shown in Supplementary Figs. 64–65, the pristine PtFe_4.7@MFI sample (named as PtFe_4.7@MFI-air) mainly comprises oxidized Pt and Fe species. After reduction treatment by H₂ at 200 °C, we can observe the presence of Pt-Pt bonding in the Pt L₃-edge EXFAS spectra, indicating the formation of metallic Pt clusters, as also confirmed by HAADF-STEM images (Supplementary Figs. 66–72). After reduction treatment at 350 °C (i.e., the reaction temperature for MCH dehydrogenation), both Pt-Fe and Pt-Pt bonding are observed. This result also infers that, even if the catalyst is pre-reduced at 200 °C by H₂, the Pt-Fe bonding will still form in the working catalyst for the MCH dehydrogenation reaction at 350 °C. Further increasing the reduction treatment temperature to 600 °C will cause a slight increase in the coordination number of Pt-Fe bonding while the coordination number of Pt-Pt bond remains almost unchanged. The evolution behaviors of the Fe and Pt species during the catalyst activation procedure are also followed by other spectroscopy and electron microscopy techniques, which confirm the gradual formation of bimetallic PtFe clusters when elevating the reduction temperature (Supplementary Figs. 62b, 73–74 and Supplementary Table 8).

Catalytic performances of PtFe@MFI catalysts

As shown in Fig. 3e, f, when measuring the catalytic performances at 400 °C and a very high WHSV of 108 h^-1 for better differentiating the catalysts, the monometallic Pt@MFI catalyst suffers a fast deactivation under this condition, which is probably caused by the aggravated coke deposition according to the characterizations of the spent catalyst (Supplementary Figs. 75–80). When the Fe loading is increased, the stability of the PtFe_x@MFI catalyst is gradually improved. With a Fe loading of 0.58 wt% (corresponding to a Fe/Pt molar ratio of 4.7), the PtFe_4.7@MFI exhibits excellent stability with a prolonged decay rate in the MCH conversion level. However, introducing an excess amount of Fe will cause sharply decreased reactivity, resulting in the loss of activity of the PtFe_7.9@MFI sample when testing at very high WHSV. The deactivation constants derived from the activity profiles show that, by incorporating an appropriate amount of Fe, the deactivation rate of the Pt@MFI catalyst can be decreased by four times (Supplementary Fig. 81).

Kinetic studies show that the bimetallic PtFe_4.7@MFI catalyst gives a lower apparent activation energy than the Pt@MFI catalyst (Supplementary Fig. 82), implying that the MCH dehydrogenation reaction is more favorable on PtFe clusters than on Pt clusters. The intrinsic reaction rates on various Pt-zeolite catalysts also infer the advantages of the bimetallic PtFe clusters over the monometallic Pt. For the bimetallic PtFe@MFI catalysts with Fe/Pt ratios ranging from 0.55 to 4.7, the more Fe the catalyst contains, the higher H₂ evolution rates it can achieve in the MCH dehydrogenation reaction. In particular, the optimal PtFe_4.7@MFI catalyst gives twice the H₂ formation rate compared to the monometallic Pt@MFI catalyst (Supplementary Fig. 83). The comparisons between the PtFe_4.7@MFI catalyst and supported PtFe catalysts prepared by other methods confirm the superior performance of the subnanometer PtFe clusters confined in MFI zeolite generated by one-pot synthesis (Supplementary Fig. 84). The exceptional performance of the PtFe_4.7@MFI catalyst is further confirmed compared to reported Pt-based catalysts regarding the specific H₂ evolution rate (Fig. 3g and Supplementary Table 9).

Considering that the H₂ evolution rate in the MCH dehydrogenation reaction strongly depends on the reaction conditions, we have carried out the thermodynamic analysis and calculations of the forward reaction rates (k_f) based on the data provided in the literature (see detailed discussion in Supplementary Note 1). The comparative analysis suggests that the PtFe_4.7@MFI catalyst developed in this work exhibits the best performance among the reported materials to date (Fig. 3h). A detailed analysis of the gaseous by-products formed in the MCH dehydrogenation reaction shows that methane is the major component with a concentration at ~500 ppm with the PtFe_4.7@MFI catalyst, corresponding to an H₂ purity of ~99.95%, which is much lower than the Pt-based catalysts supported on oxide carriers (Supplementary Fig. 85).

Theoretical studies on the Pt-based catalysts

By collective analysis of the characterization results from different techniques, we can conclude that, Pt-Fe bonding is formed in the bimetallic PtFe_x@MFI samples after reduction treatment at ≥350 °C by H₂ and the abundance of the Pt-Fe bonding depends on the Fe loading. We have employed theoretical modeling and calculations to gain insights into the structure-reactivity relationship in the bimetallic PtFe@MFI samples. Initially, we constructed three models to study the influence of the chemical composition of the Pt-based active sites on the MCH dehydrogenation reaction. Three Pt-based cluster models (Pt₇, Pt₆Sn and Pt₆Fe) made by metal clusters located in the sinusoidal 10-ring channels of MFI zeolite are constructed based on the electron microscopy characterization results to represent the Pt@MFI, PtSn_4.7@MFI and PtFe_4.7@MFI catalyst, respectively. As shown in Fig. 6a and Supplementary Fig. 86, DFT-calculated O affinity shows that both Pt₇ and Pt₆Sn clusters in the Pt₇-MFI and Pt₆Sn-MFI models tend to form metallic clusters while the Fe species tend to exist in the form of Fe(II) in the Pt₆FeO-MFI model. The results of theoretical modeling align with the spectroscopic characterization results on the chemical states of Pt and Fe species in the realistic PtFe_4.7@MFI catalysts.

**Fig. 6: Theoretical studies on methylcyclohexane dehydrogenation over Pt-based clusters confined in MFI.**

In light of the prior works on theoretical modeling of the MCH dehydrogenation process on metal catalysts, we have studied two plausible key steps in the transformation process from MCH to TOL (Supplementary Figs. 87–88)⁹. As shown in Fig. 6b, the Pt₆FeO clusters give lower energy barriers for activation of the C-H bonds in methylcyclohexene and methylcyclohexadiene than the Pt₇ and Pt₆Sn clusters, suggesting the high activity of Pt₆FeO clusters for MCH dehydrogenation reaction, which are in line with the experimental results from the kinetic studies. The advantages of Pt₆FeO clusters over the Pt₇ clusters in the activation of C-H bonds are further validated by the isotopic exchange experiments (Supplementary Fig. 89).

Prior mechanistic studies suggest that the coke deposition on the Pt catalyst and the production of methane due to hydrogenolysis of TOL can be associated with the undesired side reactions of TOL on the Pt active sites, which are responsible for the catalyst deactivation and decline in selectivity³². In this context, we have studied two elementary steps associated with the coke deposition (cleavage of the C-H bond in the methyl group of TOL) and side reaction (cleavage of the C-C bond in TOL) with the three models by DFT calculations³³. As presented in Fig. 6c and Supplementary Figs. 90–91, the Pt₇-MFI and Pt₆Sn-MFI models exhibit similar activation energy for the two elementary steps, while the Pt₆FeO-MFI model exhibits considerably higher activation energies, inferring that the undesired elementary reactions are not favorable on bimetallic PtFe clusters. Furthermore, the adsorption of toluene and *C₆H₅-CH₂ species (the coke precursor) is significantly weakened (Fig. 6d and Supplementary Fig. 92), indicating that Fe modification promotes toluene desorption and inhibits its deep dehydrogenation to produce the coke precursor, which is consistent with the activation energy analysis^34,35. Detailed analysis of the electronic structures of the three Pt-based clusters indicates that the charge transfer from the Pt atom to the adsorbed TOL molecule at the binding site will be partly decreased on Pt₆FeO-MFI model (Supplementary Fig. 93), resulting in a relatively weaker bonding interaction of TOL on the Pt₆FeO cluster in compared to the Pt₇ and Pt₆Sn cluster, as evidenced in the crystal orbital Hamilton populations (COHP) analysis (Fig. 6d, e). Consequently, the undesired transformation of TOL on Pt active sites is suppressed on the Pt₆FeO cluster (Supplementary Fig. 94).

Our experimental results show that the Fe loading and the chemical states of Fe species have marked impacts on the catalytic properties of the bimetallic PtFe_x@MFI catalysts (Supplementary Figs. 95–98). We have then studied the two elementary reactions on three different PtFe cluster models (Supplementary Fig. 99), and the results show that the incorporation of two Fe(II) atoms into the Pt clusters will cause a decrease in the energy barrier for C-H activation in TOL, which may accelerate the coke deposition on Pt sites. The results from DFT calculations are consistent with our experimental observation, in which we have found that the PtFe_7.9@MFI catalyst with a high Fe loading exhibits fast deactivation in the MCH dehydrogenation reaction. The combination of experimental and theoretical study demonstrates that the catalytic properties of subnanometer metal clusters are sensitive to chemical composition. Therefore, it is of great necessity to control the size and composition of the bimetallic clusters precisely^21,36,37.

Long-term stability tests

Long-term stability is a crucial criterion for the potential industrial implementation of the PtFe@MFI catalyst in the practical process. We first tested the stability of the optimized PtFe_4.7@MFI catalyst under the conditions in which the conversion levels of MCH are maintained at high levels (>90%). As shown in Fig. 7a, the optimized PtFe_4.7@MFI catalyst exhibits high activity and selectivity to TOL for >200 h in every single catalytic run. Moreover, the catalyst can be facilely regenerated by calcination in air and subsequent reduction by H₂ during a course of >2000 h in eight consecutive reaction-regeneration cycles. Furthermore, to confirm the full recovery of the regenerated PtFe@MFI catalyst, we have carried out consecutive reaction-regeneration tests under high WHSV conditions (i.e., maintaining the conversion levels <40%) which can better reflect the catalytic efficiency of the Pt active sites than the tests performed at high conversion levels. As shown in Fig. 7b, the activity of the deactivated PtFe@MFI catalyst can be fully recovered after calcination treatment in the air for the removal of the coke. Besides, even after elevating the calcination temperature to 600 °C, the performance of the PtFe@MFI catalyst will not be affected. Instead, the high-temperature calcination is helpful for the removal of hard coke on the spent catalyst, facilitating the recovery of the activity for MCH dehydrogenation reaction. After seven consecutive reaction-regeneration cycles for >400 h, the PtFe@MFI catalyst maintains its catalytic performance as the fresh catalyst. Structural characterizations of the spent catalysts show that the size distributions and location of the bimetallic PtFe clusters within the zeolite structure are retained after the long-term stability tests (Fig. 7c, j and Supplementary Figs. 100–107), suggesting the remarkably high activity and structural robustness of the subnanometer PtFe clusters confined in MFI zeolite (see the comparison with reported catalysts for MCH dehydrogenation reaction in Supplementary Table 9).

Discussion

In this work, we have reported a zeolite-encapsulated subnanometer PtFe catalyst for the production of high-purity H₂ from an MCH dehydrogenation reaction, which gives high activity, selectivity and stability under challenging reaction conditions. The remarkable performance is ascribed to the unique structures of the bimetallic PtFe clusters and the surrounding rigid zeolite framework. The principles for catalyst design can be extended to other H₂ storage systems based on other LOHCs (e.g., benzyltoluene and dibenzyltoluene) in which the size, composition and surrounding environment of the metal active sites within the microporous supports should be further optimized according to the molecular structures of the LOHC molecules and reaction mechanisms.

Methods

Chemicals

Zn(NO₃)₂·6H₂O (Aladdin, product code: Z190758-500g), Ga(NO₃)₂·9H₂O (Macklin, product code: G810494-1g), Mn(NO₃)₂·4H₂O (Energy Chemical, product code: E060921), In(NO₃)₂·4H₂O (Macklin, product code: I834112-25g), SnCl₄·5H₂O (Energy Chemical, product code: E010541), Fe(NO₃)₃·9H₂O (from Macklin, product code: I809398-100g), KCl (Sinopharm Chemical Reagent, product code: 10016318), H₂PtCl₆ (Adamas-beta, product code: 32606 A), Tetraethyl orthosilicate (TEOS, from Alfa Aesar, product code: 014082), Tetrapropylammonium hydroxide (TPAOH, 40 wt%, Energy Chemical, product code: 17456.22), Ethylenediamine (Alfa Aesar, product code: A12132.AP).

Synthesis of Pt-zeolite materials

Pt@MFI and a series of bimetallic PtM@MFI samples (where M represents the second metal, M=Sn, Zn, In, Ga, Fe) were synthesized by one-pot approach. In a typical synthesis procedure of PtFe_4.7@MFI sample, a TPAOH solution was initially prepared by blending 6.12 g of a 40 wt% TPAOH solution, 0.0531 g of KCl, and 14.1 g of distilled water at room temperature. Subsequently, 6.19 g of TEOS was added to TPAOH solution at 35 °C and then the suspension was kept stirring for 6 h to ensure the full hydrolysis of TEOS. Afterwards, 240 μL of H₂PtCl₆ aqueous solution (100 mg/mL), 87.5 mg of Fe(NO₃)₃ ∙ 9H₂O, and 225 μL of ethylenediamine were added to the above solution. The resulting yellow solution was then transferred to Teflon-lined autoclaves and heated in an electric oven at 175 °C for 96 h under static conditions. After the hydrothermal process, the solid product was isolated by filtration, washed with distilled water and acetone, and subsequently dried at 60 °C. The solid sample was further calcined in static air at 560 °C for 8 h with a ramp rate of 2 ^oC/min, followed by an additional heating step at 600 °C for 2 h with a ramp rate of 5 °C/min.

The bimetallic PtM@MFI catalysts (M = Sn, Zn, Ga, In and Fe) were prepared by altering the type of metal precursors while the M/Pt atomic ratio is maintained at around 4.7. Details of the synthesis procedure of PtM@MFI catalysts are provided in Supplementary Table 1. The preparation of the monometallic Pt@MFI sample was carried out in the absence of a co-catalyst metal precursor in the synthesis mixture. In addition, PtFe@MFI samples with different Fe loadings were prepared in a similar approach but the Fe loading was controlled by the mass of the Fe precursor added (Supplementary Table 2). The results of ICP-OES analysis of the chemical composition of Pt-based samples are shown in Supplementary Table 3.

Characterizations

Powder X-ray diffraction analysis was carried out with a Rigaku RU-200b X-ray powder diffractometer with Cu Kα radiation. Pore structure analysis was performed through argon adsorption-desorption on a BSD-660M A3M sorption instrument at −186 °C. Prior to the measurements, the reduced samples underwent a degassing process at 300 °C for 6 h. Specific surface area and micropore volume were determined using the Brunauer-Emmet-Teller (BET) and Saito-Foley (SF) models, respectively, while the total pore volume was estimated at P/P₀ = 0.99. The metal loading for all catalysts was quantified through inductively coupled plasma optical emission spectroscopy (ICP-OES) analysis using a Thermo IRIS Intrepid II instrument.

Magnetic susceptibility measurements were conducted using a Quantum Design SQUID VSM magnetometer. The measurements were conducted over a temperature range spanning from 300 K to 5 K and reciprocally from 5 K to 300 K, all under an external magnetic field of 100 Oe. To prevent oxidation, the sample was carefully placed into a pre-weighed SQUID capsule within a glovebox. Subsequently, the capsule was sealed, weighed, and transferred to the SQUID cavity for magnetic susceptibility measurements.

The quasi-in situ ⁵⁷Fe Mössbauer spectra were acquired at 80 K using an MFD-500AV spectrometer (Topologic Systems, Japan) with a proportional counter, utilizing ⁵⁷Co(Rh) as the γ-ray radioactive source. Before characterization, the sample was reduced in a 5% H₂/Ar atmosphere for 3 h at temperatures of 600 °C. After the pre-reduction treatment, the sample was carefully transferred into a capsule within a glovebox to avoid contact with air. The measuring velocity was calibrated using a standard α-iron foil. The data analysis was performed with Moss Winn 4.0i software.

UV-visible diffuse reflectance spectroscopy was performed using an ultraviolet spectrophotometer (Lambda 950) within the wavelength range of 200–800 nm. UV-Raman spectroscopic characterization was conducted at room temperature using the HORIBA JY LabRAM HR Evolution Raman spectrometer with a 266 nm excitation source.

The number of exposed metal sites in each Pt-zeolite catalyst was determined by CO pulse chemisorption experiments, which were carried out on an automatic chemisorption analyzer (BSD-Chem C200). The solid samples (100–200 mg) were subjected to pretreatment under an H₂ atmosphere at a given temperature for 3 h, followed by cooling to the adsorption temperature at ~50 °C. After purging with helium, CO pulses were injected multiple times until the cumulative quantity of adsorbed CO reached saturation. The quantity adsorbed for each CO pulse was recorded using a thermal conductivity detector, and the metal dispersion was calculated assuming a ratio of surface Pt/adsorbed CO equal to 1. The equations used for the calculation of metal dispersion (D) are provided below:

$${D}\left(\%\right)=\frac{{{N}}_{{m}}{{F}}_{{s}}{M}\times 1{0}^{4}}{{L}}$$

(1)

Where N_m represents the chemisorption uptake expressed in mol of CO per gram of the sample, and F_s represents the adsorption stoichiometry, assumed to be 1 in our measurement. M represents the molecular weight of the supported metal (Pt), and L represents the loading of Pt.

In situ CO diffuse reflectance infrared Fourier transform spectra (CO-DRIFTS) were acquired using a Bruker Tensor 27 spectrometer equipped with a mercury cadmium telluride (MCT) detector cooled by liquid nitrogen. Before the measurements, the sample was subjected to a H₂ reduction treatment at a given temperature for 3 h. Following the reduction pretreatment, the sample was exposed to 10 vol% CO/N₂ at 30 °C for 30–50 min, and the spectra were collected after purging with N₂ for an additional 30–50 min.

Samples for electron microscopy studies were prepared by dropping the suspension of the solid samples in CH₂Cl₂ directly onto holey-carbon coated copper grids. Electron Microscopy measurements were performed using two types of microscopes. A JEOL 2100 F microscope operating at 200 kV was used to record HAADF-STEM images at low resolution. High-resolution HAADF-STEM and iDPC-STEM imaging was performed on a FEI Titan³ Themis 60–300 microscope working at 300 kV. To reduce the electron beam damage to the sample, the images were recoded with a beam current of 10–30 pA and a 2.5 μs dwell time. More details about the optimization of the measurement conditions of Pt-zeolite materials can be referred to prior work^19,38.

X-ray absorption experiments at the Pt (11,564 eV) L_III and Fe (7112 eV) K edges were performed in NOTOS beamline in ALBA synchrotron. The beam was monochromatized using Si(111) double crystals; harmonic rejection was performed setting the correct angle of collimating and focusing mirrors. Si and Rh stripes have been chosen to adjust the reflectivity curves. The spectra were collected in fluorescence modes by means of the ionization chambers filled with appropriate gases (95% N₂ + 5% Kr for I₀ (ionization chamber 0, located before the sample) and 17.1% N₂ + 82.9% Kr for I₁ (ionization chamber 1, located after the sample)), and a fluorescence solid-state detector. Samples was diluted using BN and loaded inside a high-pressure gas cell with one end of the both-way open-ended quartz capillary (diameter of 1 mm) is connected to desired gas, and the other end is connected to the mass spectrometer for on-line gas analysis. Several scans were acquired at each measurement step to ensure spectral reproducibility and good signal-to-noise ratio. The data reduction and extraction of the χ(k) function were performed using the Athena code. EXAFS data analysis was performed using the Artemis software. Phase and amplitudes were calculated with the FEFF6 code.

Theoretical calculations

Density functional theory calculations were completed via Vienna ab initio simulation software package with the atomic simulation environment^39,40. As a form of the generalized gradient approximation, the Perdew−Burke−Ernzerhof functional was carried out to describe the electron exchange and correlation effects⁴¹. DFT-D3 method with Becke-Jonson damping was utilized to deal with van der Waals correction for the catalytic systems⁴². The projector-augmented wave method was used to treat the interaction between atomic cores and electrons⁴³. The plane-wave basis set was employed with a cutoff energy of 400 eV. Besides, the atomic force convergence criterion was set to 0.03 eV/Å. Three Pt-based clusters (Pt₇, Pt₆Sn and Pt₆Fe) confined in MFI zeolite were constructed to represent the Pt@MFI, PtSn@MFI and PtFe@MFI catalysts, respectively. A 1 × 1 × 1 Gamma point was selected for geometry optimization. Besides, all models were fully relaxed.

To find transition state geometries, a combination of the climbing image nudged elastic band (CI-NEB) method and dimer method was employed with a total of 6 images^44,45. There was only one imaginary frequency for all transition state structures.

The adsorption energy of the adsorbate is calculated by the following equation:

$$E\left(*{\mbox{ads}}\right)=E\left({\mbox{clean}}+{\mbox{ads}}\right)-E\left({\mbox{clean}}\right)-E\left({\mbox{gas}}\right)$$

(2)

Where E(clean+ads) is the energy for the model after the adsorption of the adsorbate, E(clean) is the energy for the model without any adsorbate, E(gas) is the energy for the adsorbate in its gas phase.

Here, we took the energies of methylcyclohexane(g) and hydrogen(g) as references, the adsorption energy of other species was further calculated by the reference energies. Take methylbenzene as an example:

$$E\left(*{{\mbox{C}}}_{7}{{\mbox{H}}}_{8}\right)=E\left({\mbox{clean}}+{{\mbox{C}}}_{7}{{\mbox{H}}}_{8}\right)-E\left({\mbox{clean}}\right)-[E({{\mbox{C}}}_{7}{{\mbox{H}}}_{14\left({\mbox{g}}\right)})-3E({{\mbox{H}}}_{2\left({\mbox{g}}\right)})]$$

(3)

Where E($*$C₇H₈) is the adsorption energy of methylbenzene species on the model, E(C₇H_14(g)) and E(H₂) are the reference energies.

The activation energy of an elementary step is calculated by the following equations:

$${{E}}_{{{\rm{act}}}}={{E}}_{{{\rm{TS}}}}-{{E}}_{{\mbox{IS}}}$$

(4)

Where E_act is the activation energy of a specific elementary step. E_IS and E_TS are initial state energy and transition state energy for the step, respectively.

Catalytic testing

For screening the candidate materials for MCH dehydrogenation reaction, the catalytic tests were conducted in a fixed-bed reactor under atmospheric pressure using a mixture of MCH/N₂ as the feed gas at 350 °C. Prior to the reaction, the PtM@MFI catalyst was pre-reduced by H₂ (5% H₂ diluted by Ar, 20 mL/min) at 600 °C for 3 h with a ramp rate of 10 °C/min. Following the reduction pre-treatment, the sample was cooled to 350 °C in a N₂ atmosphere. Subsequently, the reaction was initiated by introducing the reactant mixture (MCH:N₂ = 1:1). The N₂ feed was regulated by a mass flow controller, and the MCH feed was controlled using an injection pump.

When measuring the initial reaction rates, a fresh Pt-zeolite catalyst was used for each measurement at a specific temperature. By using a mixture of MCH and H₂ as the reaction feed, at the initial several hours, the Pt-zeolite catalyst remains relatively stable, allowing us to acquire a reliable initial reaction rate for calculating the specific reaction rates normalized to the total mass of Pt species in the solid zeolite or to the exposed Pt sites determined by CO chemisorption.

To investigate the apparent activation energy of a series of catalysts, the conversions of MCH were adjusted to below 25% by varying the weight-hour-space-velocity (WHSV) while the feed composition is fixed to MCH/H₂ mixture with an MCH/H₂ of 1/1. For obtaining the apparent activation energy, the yield of TOL was measured at 300, 320, 350, 380, and 400 °C.

To expedite the catalyst selection process, MCH conversion levels were adjusted the weight-hour-space-velocity (WHSV) while maintaining a constant ratio of MCH/N₂ in the feed gas. The products were analysed using an online GC (Agilent 8890 equipped with TCD and FID detectors). The conversion of MCH (X_MCH), selectivity of TOL (S_TOL), yield of TOL (Y_TOL), H₂ evolution rate (${R}_{{{\mbox{H}}}_{2}}$), and the deactivation constant (k_d) were calculated based on the following Equations.

$${X}_{{\mbox{MCH}}}=\frac{{F}_{{\mbox{MCH}},{\mbox{in}}}-{F}_{{\mbox{MCH}},{\mbox{out}}}}{{F}_{{\mbox{MCH}},{\mbox{in}}}}\times 100\%$$

(5)

$${S}_{{\mbox{TOL}}}=\frac{{F}_{{\mbox{TOL}},{\mbox{out}}}}{{F}_{{\mbox{MCH}},{\mbox{in}}}-{F}_{{\mbox{MCH}},{\mbox{out}}}}\times 100\%$$

(6)

$${{\mbox{Y}}}_{{\mbox{TOL}}}={{\mbox{X}}}_{{\mbox{MCH}}}\times {{\mbox{S}}}_{{\mbox{TOL}}}\times 100\%$$

(7)

$${R}_{{{\mbox{H}}}_{2}}\left({\mbox{mmol}}/{{\mbox{g}}}_{{\mbox{Pt}}}/\min \right)=\frac{3\times {Y}_{{\mbox{TOL}}}\times {F}_{{\mbox{MCH}},{\mbox{in}}}}{{{\mbox{M}}}_{{\mbox{Pt}}}}$$

(8)

$${{\mbox{k}}}_{{\mbox{d}}}\left({{\mbox{h}}}^{-1}\right)=\frac{{{\mathrm{ln}}}\frac{1-{{\mbox{X}}}_{{\mbox{final}}}}{{{\mbox{X}}}_{{\mbox{final}}}}-{{\mathrm{ln}}}\frac{1-{{\mbox{X}}}_{{\mbox{initial}}}}{{{\mbox{X}}}_{{\mbox{initial}}}}}{{\mbox{t}}}$$

(9)

where F_MCH,in and F_MCH,out, represent the flow velocity of MCH in the feed and effluent, respectively and F_TOL,out represents the flow velocity of TOL in the effluent; M_Pt represents the mass of Pt in the solid catalyst. X_initial and X_final represent the initial and final conversion of MCH, respectively, and t represents the time the reaction time.

Long-term stability tests

To assess the catalyst’s regeneration stability under severe deactivation conditions and working stability under high conversion conditions, the conversion levels of MCH were kept below 50% to facilitate the detection of any potential deactivation of the varying the PtFe_4.8@MFI catalyst during consecutive reaction-regeneration cycles. To simulate MCH dehydrogenation under practical working conditions, the reactant feed was changed to MCH/H₂ mixture with an MCH/H₂ of 1/1 which allows the downstream application of the H₂ released from LOHC. Before the catalytic test, the PtFe_4.8@MFI catalyst was subjected to a pre-reduction treatment by H₂ (5% H₂ diluted by Ar, 20 mL/min) at 200 °C for 3 h. Following the reduction pre-treatment, the catalytic test was initiated by introducing the reactant gas mixture (MCH:H₂ = 1:1) under atmospheric pressure at 350 °C with a WHSV of 43 h^-1, with a ramp rate of 10 °C/min from 200 to 350 °C. After the initial MCH dehydrogenation test, the sample was cooled to ~100 °C and initiated the regeneration process through calcination in air at 450 °C for 3 h, with a ramp rate of 2 °C/min. Following air calcination, the sample was cooled to 100 °C and then subjected to reduction with H₂ at 200 °C for 3 h. After reduction, the second cycle of the MCH dehydrogenation reaction was initiated by introducing the reactant gas mixture (MCH:H₂ = 1:1) under atmospheric pressure at 350 °C, with a ramp rate of 10 °C/min from 200 to 350 °C. The regeneration procedure for the samples after the third and fifth cycles remained consistent with the process after the first MCH dehydrogenation. However, for the regeneration of the samples after the fourth and sixth cycles of the MCH dehydrogenation reaction, the calcination temperature was increased to 600 °C, while the other regeneration procedures remained the same as for the samples after the first MCH dehydrogenation.

To evaluate the performance of the PtFe_4.8@MFI catalyst at high MCH conversion levels, the WHSV is changed to 4.3 h^-1 at atmosphere pressure while the reaction temperature is maintained at 350 °C. The pretreatment and the regeneration procedure is the same as the stability tests performed at high WHSV.

Data availability

The data that support the findings of this study are present in the paper and/or the Supplementary Information and are available from the corresponding authors on request.

References

Nazir, H. et al. Is the H₂ economy realizable in the foreseeable future? Part II: H₂ storage, transportation, and distribution. Int. J. Hydrog. Energy 45, 20693–20708 (2020).
Article ADS CAS MATH Google Scholar
Abdalla, A. M. et al. Hydrogen production, storage, transportation and key challenges with applications: a review. Energy Convers. Manag. 165, 602–627 (2018).
Article ADS CAS MATH Google Scholar
Preuster, P., Papp, C. & Wasserscheid, P. Liquid Organic Hydrogen Carriers (LOHCs): toward a hydrogen-free hydrogen economy. Acc. Chem. Res. 50, 74–85 (2016).
Article PubMed Google Scholar
Wei, D. et al. Toward a hydrogen economy: development of heterogeneous catalysts for chemical hydrogen storage and release reactions. ACS Energy Lett. 7, 3734–3752 (2022).
Article CAS MATH Google Scholar
Salman, M. S., Rambhujun, N., Pratthana, C., Srivastava, K. & Aguey-Zinsou, K.-F. Catalysis in liquid organic hydrogen storage: recent advances, challenges, and perspectives. Ind. Eng. Chem. Res. 61, 6067–6105 (2022).
Article CAS Google Scholar
Niermann, M., Timmerberg, S., Drünert, S. & Kaltschmitt, M. Liquid organic hydrogen carriers and alternatives for international transport of renewable hydrogen. Renew. Sustain. Energy Rev. 135, 110171 (2021).
Article CAS Google Scholar
Liu, L. et al. Atomic-level understanding on the evolution behavior of subnanometric Pt and Sn species during high-temperature treatments for generation of dense PtSn clusters in zeolites. J. Catal. 391, 11–24 (2020).
Article CAS Google Scholar
Gianotti, E., Taillades-Jacquin, M., Rozière, J. & Jones, D. J. High-purity hydrogen generation via dehydrogenation of organic carriers: a review on the catalytic process. ACS Catal. 8, 4660–4680 (2018).
Article CAS Google Scholar
Zhao, W., Chizallet, C., Sautet, P. & Raybaud, P. Dehydrogenation mechanisms of methyl-cyclohexane on γ-Al₂O₃ supported Pt₁₃: impact of cluster ductility. J. Catal. 370, 118–129 (2019).
Article CAS Google Scholar
Bulgarin, A., Jorschick, H., Preuster, P., Bösmann, A. & Wasserscheid, P. Purity of hydrogen released from the liquid organic hydrogen carrier compound perhydro dibenzyltoluene by catalytic dehydrogenation. Int. J. Hydrog. Energy 45, 712–720 (2020).
Article ADS CAS Google Scholar
Nakaya, Y., Miyazaki, M., Yamazoe, S., Shimizu, K. I. & Furukawa, S. Active, selective, and durable catalyst for alkane dehydrogenation based on a well-designed trimetallic alloy. ACS Catal. 10, 5163–5172 (2020).
Article CAS Google Scholar
Chen, L. et al. Reversible dehydrogenation and rehydrogenation of cyclohexane and methylcyclohexane by single-site platinum catalyst. Nat. Commun. 13, 1092 (2022).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Wang, J. et al. Size-dependent catalytic cyclohexane dehydrogenation with platinum nanoparticles on nitrogen-doped carbon. Energy Fuels 34, 16542–16551 (2020).
Article CAS MATH Google Scholar
Al-ShaikhAli, A. H., Jedidi, A., Anjum, D. H., Cavallo, L. & Takanabe, K. Kinetics on NiZn bimetallic catalysts for hydrogen evolution via selective dehydrogenation of methylcyclohexane to toluene. ACS Catal. 7, 1592–1600 (2017).
Article CAS Google Scholar
Murata, K. et al. Structure–property relationships of Pt–Sn nanoparticles supported on Al₂O₃ for the dehydrogenation of methylcyclohexane. ACS Appl. Nano Mater. 4, 4532–4541 (2021).
Article CAS MATH Google Scholar
Yoshimi, O. et al. SPERA hydrogen systemtm by LOHC-MCH method for massive hydrogen storage and transportation. J. Environ. Sci. Eng. A 13, 50–60 (2024).
MATH Google Scholar
Alhumaidan, F., Cresswell, D. & Garforth, A. Hydrogen storage in liquid organic hydride: producing hydrogen catalytically from methylcyclohexane. Energy Fuels 25, 4217–4234 (2011).
Article CAS MATH Google Scholar
Liu, L. & Corma, A. Confining isolated atoms and clusters in crystalline porous materials for catalysis. Nat. Rev. Mater. 6, 244–263 (2020).
Article ADS MATH Google Scholar
Liu, L. et al. Regioselective generation and reactivity control of subnanometric platinum clusters in zeolites for high-temperature catalysis. Nat. Mater. 18, 866–873 (2019).
Article ADS CAS PubMed MATH Google Scholar
Liu, L. & Corma, A. Identification of the active sites in supported subnanometric metal catalysts. Nat. Catal. 4, 453–456 (2021).
Article CAS MATH Google Scholar
Liu, L. & Corma, A. Bimetallic sites for catalysis: from binuclear metal sites to bimetallic nanoclusters and nanoparticles. Chem. Rev. 123, 4855–4933 (2023).
Article CAS PubMed PubMed Central MATH Google Scholar
Nakaya, Y. & Furukawa, S. Catalysis of alloys: classification, principles, and design for a variety of materials and reactions. Chem. Rev. 123, 5859–5947 (2022).
Article PubMed MATH Google Scholar
Liu, L., Lopez-Haro, M., Calvino, J. J. & Corma, A. Tutorial: structural characterization of isolated metal atoms and subnanometric metal clusters in zeolites. Nat. Protoc. 16, 1871–1906 (2020).
Article PubMed Google Scholar
Biswas, S. et al. Origin of regioselectivity in the dehydrogenation of alkanes by pincer–iridium complexes: a combined experimental and computational study. ACS Catal. 11, 12038–12051 (2021).
Article CAS Google Scholar
Jentys, A. Estimation of mean size and shape of small metal particles by EXAFS. Phys. Chem. Chem. Phys. 1, 4059–4063 (1999).
Article CAS MATH Google Scholar
Liu, L. et al. Structural modulation and direct measurement of subnanometric bimetallic PtSn clusters confined in zeolites. Nat. Catal. 3, 628–638 (2020).
Article CAS MATH Google Scholar
Liu, Y. et al. Isomorphously Substituted [Fe,Al]ZSM-5 Catalysts for Methane Dehydroaromatization. ACS Catal. 13, 8128–8138 (2023).
Article CAS MATH Google Scholar
Boubnov, A. et al. Selective catalytic reduction of NO Over Fe-ZSM-5: mechanistic insights by Operando HERFD-XANES and Valence-to-Core X-ray emission spectroscopy. J. Am. Chem. Soc. 136, 13006–13015 (2014).
Article CAS PubMed MATH Google Scholar
Buttignol, F. et al. Aging of industrial Fe-zeolite based catalysts for nitrous oxide abatement in nitric acid production plants. Catal. Sci. Technol. 12, 7308–7321 (2022).
Article CAS MATH Google Scholar
Kornas, A., Mlekodaj, K. & Tabor, E. Nature and redox properties of iron sites in zeolites revealed by Mössbauer spectroscopy. ChemPlusChem 89, e202300543 (2023).
Article PubMed Google Scholar
Agote-Arán, M. et al. Operando HERFD-XANES/XES studies reveal differences in the activity of Fe-species in MFI and CHA structures for the standard selective catalytic reduction of NO with NH3. Appl. Catal. A Gen. 570, 283–291 (2019).
Article Google Scholar
Antunes, A. P. et al. Catalytic oxidation of toluene over CuNaHY zeolites. Appl. Catal. B Environ. 33, 149–164 (2001).
Article CAS Google Scholar
Du, S., Gamliel, D. P., Giotto, M. V., Valla, J. A. & Bollas, G. M. Coke formation of model compounds relevant to pyrolysis bio-oil over ZSM-5. Appl. Catal. A Gen. 513, 67–81 (2016).
Article CAS Google Scholar
Furukawa, S. et al. Remarkable enhancement in hydrogenation ability by phosphidation of ruthenium: specific surface structure having unique Ru ensembles. ACS Catal. 8, 8177–8181 (2018).
Article CAS MATH Google Scholar
Tsai, M. C. & Muetterties, E. L. Platinum metal surface chemistry of benzene and toluene. J. Am. Chem. Soc. 104, 2534–2539 (2002).
Article MATH Google Scholar
Alexeev, O. S. & Gates, B. C. Supported bimetallic cluster catalysts. Ind. Eng. Chem. Res. 42, 1571–1587 (2002).
Article Google Scholar
Lavroff, R. H., Morgan, HarryW. T., Zhang, Z., Poths, P. & Alexandrova, A. N. Ensemble representation of catalytic interfaces: soloists, orchestras, and everything in-between. Chem. Sci. 13, 8003–8016 (2022).
Article CAS PubMed PubMed Central MATH Google Scholar
Liu, L., Lopez-Haro, M., Calvino, J. J. & Corma, A. Tutorial: structural characterization of isolated metal atoms and subnanometric metal clusters in zeolites. Nat. Protoc. 16, 1871–1906 (2021).
Article CAS PubMed Google Scholar
Kresse, G. & Furthmüller, J. Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set. Comput. Mater. Sci. 6, 15–50 (1996).
Article CAS MATH Google Scholar
Kresse, G. & Hafner, J. Ab initiomolecular-dynamics simulation of the liquid-metal–amorphous-semiconductor transition in germanium. Phys. Rev. B 49, 14251–14269 (1994).
Article ADS CAS MATH Google Scholar
Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865–3868 (1996).
Article ADS CAS PubMed Google Scholar
Grimme, S., Ehrlich, S. & Goerigk, L. Effect of the damping function in dispersion corrected density functional theory. J. Comput.l Chem. 32, 1456–1465 (2011).
Article CAS MATH Google Scholar
Kresse, G. & Joubert, D. From ultrasoft pseudopotentials to the projector augmented-wave method. Phys. Rev. B 59, 1758–1775 (1999).
Article ADS CAS MATH Google Scholar
Henkelman, G. & Jónsson, H. Improved tangent estimate in the nudged elastic band method for finding minimum energy paths and saddle points. J. Chem. Phys. 113, 9978–9985 (2000).
Article ADS CAS MATH Google Scholar
Henkelman, G. & Jónsson, H. A dimer method for finding saddle points on high dimensional potential surfaces using only first derivatives. J. Chem. Phys. 111, 7010–7022 (1999).
Article ADS CAS MATH Google Scholar

Download references

Acknowledgements

This study is sponsored by Tsinghua-Toyota Joint Research Institute Inter-disciplinary Program. The in situ XAS experiments were performed at NOTOS BL16 beamline at ALBA Synchrotron with the collaboration of ALBA staff. We thank Dr. R. Zong for his help in the electron microscopy characterization of the Pt-zeolite catalysts at Tsinghua University.

Author information

These authors contributed equally: Zhe He, Kailang Li, Tianxiang Chen.

Authors and Affiliations

Engineering Research Center of Advanced Rare-Earth Materials of Ministry of Education, Department of Chemistry, Tsinghua University, Beijing, China
Zhe He, Yunchao Feng & Lichen Liu
Center for Combustion Energy, Tsinghua University, Beijing, China
Kailang Li & Liang Zhang
School of Vehicle and Mobility, Tsinghua University, Beijing, China
Kailang Li, Fuyuan Yang & Liang Zhang
Department of Applied Biology and Chemical Technology, The Hong Kong Polytechnic University, Hunghom, Hong Kong, China
Tianxiang Chen & Tsz Woon Benedict Lo
ALBA Synchrotron Light Source, Cerdanyola del Vallès, Barcelona, Spain
Eduardo Villalobos-Portillo & Carlo Marini

Authors

Zhe He
View author publications
Search author on:PubMed Google Scholar
Kailang Li
View author publications
Search author on:PubMed Google Scholar
Tianxiang Chen
View author publications
Search author on:PubMed Google Scholar
Yunchao Feng
View author publications
Search author on:PubMed Google Scholar
Eduardo Villalobos-Portillo
View author publications
Search author on:PubMed Google Scholar
Carlo Marini
View author publications
Search author on:PubMed Google Scholar
Tsz Woon Benedict Lo
View author publications
Search author on:PubMed Google Scholar
Fuyuan Yang
View author publications
Search author on:PubMed Google Scholar
Liang Zhang
View author publications
Search author on:PubMed Google Scholar
Lichen Liu
View author publications
Search author on:PubMed Google Scholar

Contributions

L.L. conceived the project. L.Z. designed theoretical studies. Z.H. carried out the materials synthesis and the structural characterizations. K.L. performed the theoretical calculations. T.C. and B.T.W.L. conducted the in situ XAS measurements and analysed the data. Y.F. contributed to the STEM characterization of the Pt-zeolite materials. E.V. and C.M. contributed to the spectroscopic characterization of Pt-zeolite materials by XAS. F.Y. contributed to the analysis of the catalytic results. L.L. wrote the manuscript with input from all the other authors.

Corresponding authors

Correspondence to Tsz Woon Benedict Lo, Fuyuan Yang, Liang Zhang or Lichen Liu.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Gerhard Pirngruber and Kazuhiro Takanabe for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information (download PDF )

Transparent Peer Review file (download PDF )

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

He, Z., Li, K., Chen, T. et al. High-purity hydrogen production from dehydrogenation of methylcyclohexane catalyzed by zeolite-encapsulated subnanometer platinum-iron clusters. Nat Commun 16, 92 (2025). https://doi.org/10.1038/s41467-024-55370-z

Download citation

Received: 20 May 2024
Accepted: 09 December 2024
Published: 02 January 2025
Version of record: 02 January 2025
DOI: https://doi.org/10.1038/s41467-024-55370-z

Subjects

Abstract

Similar content being viewed by others

Methyl formate as a hydrogen energy carrier

Catalytic acceptorless complete dehydrogenation of cycloalkanes

Achieving gigawatt-scale green hydrogen production and seasonal storage at industrial locations across the U.S

Introduction

Results

Synthesis of Pt-based catalysts

Evaluation of the Pt-based catalysts for MCH dehydrogenation

Structural characterizations of the PtFe@MFI catalysts

Catalytic performances of PtFe@MFI catalysts

Theoretical studies on the Pt-based catalysts

Long-term stability tests

Discussion

Methods

Chemicals

Synthesis of Pt-zeolite materials

Characterizations

Theoretical calculations

Catalytic testing

Long-term stability tests

Data availability

References

Acknowledgements

Author information

Authors and Affiliations

Contributions

Corresponding authors

Ethics declarations

Competing interests

Peer review

Peer review information

Additional information

Supplementary information

Supplementary Information (download PDF )

Transparent Peer Review file (download PDF )

Rights and permissions

About this article

Cite this article

Share this article

Search

Quick links