Stabilizing Pt–Fe dual-metal single atoms in ZIFs: a pathway to form heterogeneous catalysts

Abstract

Dual-atom catalysts (DACs) show promise for enhanced catalytic performance through synergistic metal interactions, yet their formation and stability during high-temperature pyrolysis remain poorly understood. In this study, we report the atomic-scale structural evolution and stabilization of Pt–Fe hetero-pairs anchored on N-doped carbon derived from zeolitic imidazolate framework-8 (ZIF-8) using in situ high-resolution high-angle annular dark-field scanning transmission electron microscopy (HR HAADF-STEM) during pyrolysis up to 900 °C. Post-pyrolysis analysis of Pt(acac)₂/Fe(acac)₃-encapsulated ZIF (denoted as PF-ZIF) at 900 °C confirms the formation of stable Pt–Fe hetero-pairs, driven by enhanced electronic coupling with the N-doped carbon matrix. In contrast, the Fe atoms in Fe(acac)₃-encapsulated ZIF (denoted as F-ZIF), tend to aggregate into Fe₃C nanoparticles (NPs) under electron beam exposure at 500 °C, rather than remaining as Fe single atoms (SAs). Additionally, increasing Pt and Fe precursor concentrations in PF-ZIF (denoted as HPF-ZIF) drives a transition from SAs to amorphous nanoclusters (NCs), culminating in crystalline Pt-doped α-Fe NPs, highlighting robust Pt–Fe bonding and temperature-dependent phase transitions. Our findings empower the precise tailoring of synthesis strategies, offer insights into atomic-scale structural evolution mechanisms, and establish fundamental design principles for advancing dual-metal materials.

---

New concepts

This study demonstrates a breakthrough in stabilization of Pt–Fe dual-metal single atoms (SAs) within zeolitic imidazolate framework (ZIF)-derived N-doped carbon during high-temperature pyrolysis up to 900 °C, using in situ high-resolution HAADF-STEM to reveal atomic-scale structural evolution. Unlike prior research studies focusing on single-atom catalysts (SACs), which often suffer from aggregation into nanoparticles, our work elucidates the formation and stability of Pt–Fe hetero-pairs, driven by enhanced electronic coupling with the N-doped carbon matrix. This differentiates our approach from existing studies that primarily report catalytic performance without detailing dynamic structural transformations. The observed stability of Pt–Fe hetero-pairs, contrasting with Fe SA aggregation into Fe₃C nanoclusters in single-metal systems, highlights the critical role of metal–metal interactions in preventing evaporation and clustering. Additionally, our findings reveal a concentration-dependent transition from SAs to Pt-doped α-Fe nanoparticles in high-precursor systems, offering a mechanistic framework for tailoring dual-metal catalysts. This work provides new insights into designing stable, high-performance dual-atom catalysts (DACs), advancing materials science by establishing synthesis strategies and atomic-level understanding for next-generation heterogeneous catalysts with applications in energy and chemical industries.

Introduction

Heterogeneous catalysts are indispensable to modern chemical industries,¹ where transition or noble metals are commonly immobilized on solid phases to enhance catalytic efficiency through the dispersion of ultra-small metal particles.² Recently, the emergence of single-atom catalysts (SACs) has attracted significant attention for dispersing metal atoms on supports at an atomic level, maximizing metal utilization and ensuring uniform active sites.^3,4 However, challenges such as high surface energy often result in the coexistence of metal particles or clusters alongside single atoms (SAs), thereby limiting stable dispersion.^5–7 Therefore, developing methodologies that promote atomic dispersion while preserving stability is crucial. In recent years, multiple strategies have been employed to achieve stable dispersion of SACs. The bottom-up approach involves anchoring metal ions within defects of solid matrices (e.g., zeolites,^8,9 metal–organic frameworks,^10,11 or carbon-based materials^12,13), effectively preventing aggregation into nanoparticles (NPs). Recent advancements highlight high-temperatures as being effective in converting noble metal NPs into SAs,¹⁴ whereas alternative low-temperature methods remain underexplored due to thermodynamic constraints. Furthermore, strategies aimed at increasing single atom density on substrates while stabilizing atoms have led to the development of dual-atom catalysts (DACs),¹⁵ comprising adjacent metal atoms that synergistically modulate activation and adsorption energies of reactants,¹⁶ thereby extending and enhancing the functionality of SACs.¹⁷ For instance, Han et al. reported that Fe-N₄/Pt-N₄@NC exhibited superior ORR activity and selectivity due to the optimized electronic structure and the synergistic effect between adjacent metal centers, which enhances both O₂ activation and adsorption.¹⁶ Although numerous studies have reported the catalytic performance of DACs,^16,18–38 most of them focus on activity metrics without fully elucidating the dynamic structural evolution of the metal centers during synthesis. A deeper understanding of these transformations is essential for rational catalyst design. In situ transmission electron microscopy (TEM), particularly scanning TEM (STEM), is a pivotal tool for elucidating the formation mechanisms and material transformations of single-atom materials at the atomic scale. By enabling real-time monitoring of dynamic processes—such as nucleation, growth, and stability—under operational conditions,^39,40in situ TEM provides critical insights into the atomic-scale behavior of SACs, DACs, and related systems.^14,41,42 Furthermore, high-resolution, aberration-corrected STEM achieves sub-Ångström resolution, revealing atomic arrangements, defect dynamics, and phase transitions under relevant temperatures, pressures, or electron beam conditions.^43–45 Recent advances in operando investigations have further expanded this field. For instance, Wang et al. reported the structural evolution of Ir and Ni SAs within ZIF-derived porous Co/N–C by in situ STEM.⁴⁶ In addition, Liu et al. directly monitored RuCu active sites and intermediates under working conditions using operando spectroscopy combined with DFT calculations.⁴⁷ Although these studies offer valuable mechanistic insights, direct atomic-scale observations of DAC structural evolution remain scarce. Addressing this gap is essential for optimizing synthesis strategies, understanding structure–property relationships, and advancing the functional performance of these materials in catalytic, electronic, and energy applications.

Herein, we investigated the structural evolution of Pt and Fe atoms as dual-metals within a zeolitic imidazolate framework (ZIF) during high-temperature pyrolysis. The synthesis of metal-encapsulated ZIFs is facile, scalable, and cost-effective, yielding materials with uniformly dispersed metal species and tunable structures. Upon pyrolysis, these ZIFs transform into porous, high-surface-area N-doped carbon. Furthermore, the nitrogen defects thus formed not only serve as active sites for various catalytic reactions, but more critically, provide coordination environments that robustly stabilize metal SAs via M–N_x bonding. In situ STEM observations revealed that, after the pyrolysis of Pt(acac)₂/Fe(acac)₃-encapsulated ZIF (denoted as PF-ZIF) at 900 °C, stable Pt–Fe hetero-pairs formed and persisted, highlighting the role of electronic coupling with the N-doped carbon matrix. In contrast, in the pyrolysis of Fe(acac)₃-encapsulated ZIF (denoted as F-ZIF), a transformation from SAs into amorphous nanoclusters (NCs), followed by the formation of crystalline Fe₃C NPs under electron beam irradiation, was observed. These atomic-level observations enable precise control of synthesis methods, deepen understanding of the structural evolution mechanism, and elucidate design principles for dual-metal system materials.

Results and discussion

Structural characterization of Pt–Fe hetero-pairs in F-900

Initially, Fe(acac)₃ was encapsulated as the Fe precursor in the ZIF (denoted as F-ZIF) during synthesis. Subsequently, Fe single atoms (SAs) were successfully formed by pyrolyzing F-ZIF at 900 °C [denoted as F-900]. Similarly, equimolar ratios of Pt(acac)₂ and Fe(acac)₃ precursors were simultaneously added into the solution during ZIF-8 synthesis process (denoted as PF-ZIF), Pt–Fe dual-metal SAs were then obtained after pyrolysis [Fig. 1(a)]. Fig. S1 and S2(a–c) are XRD patterns, SEM and TEM images of PF-ZIF, the diffraction peaks correspond to ZIF-8 structures, SEM and TEM images reveal the rhombic dodecahedron morphology of the ZIF-8 with the size ranging from 100 to 200 nm. By employing the cage confinement strategies, both Pt and Fe metal precursors were encapsulated within the pores without changing the ZIF-8 structure [Fig. S3]. ICP-OES analysis confirmed the incorporation of Pt and Fe, with respective loadings of 0.0403 wt% and 0.0612 wt%, thereby verifying the effective encapsulation of both metal precursors within the porous framework. Upon pyrolysis of PF-ZIF crystals at 900 °C [denoted as PF-900], the framework was transformed into graphite. The overall morphology remains unchanged, while a slight reduction and a significant increase in surface roughness is observed, as shown in Fig. 1(b and c) and Fig. S2(a–c). Additionally, Fig. 1(d and e) show the TEM and HAADF-STEM images of the PF-900; obviously, no NPs or clusters were found in the matrix. Notably, high-resolution (HR) HAADF-STEM images in Fig. 1(f) disclose many bright spots with different contrasts distributed in PF-900. It is known that HAADF-STEM imaging provides the atomic number-dependent contrast; the elements with larger atomic number have brighter contrast for the imaging.^48,49 Due to the significant disparity in atomic numbers between Pt and Fe, Fe atoms appear as darker spots while Pt atoms appear brighter in HAADF-STEM imaging. In addition to the individually dispersed SAs, several Pt–Fe dual-metal hetero-pairs are also observed in the STEM image [Fig. 1(f)]. Furthermore, Fig. 1(g) highlights the selected Pt–Fe hetero-pairs extracted from Fig. 1(f), revealing a consistent interatomic distance between adjacent metal atoms. The intensity profile in Fig. 1(h), corresponding to site #1 labeled in Fig. 1(f), clearly demonstrates the intensity contrast of two neighboring metal atoms. The spots with high and low intensity attributed to Pt and Fe atoms, respectively. Additionally, the measured interatomic distance of 0.53 nm between adjacent metal atoms matches the most stable Pt–Fe structure reported in a previous study.¹⁶ STEM image and corresponding EDS elemental mappings of PF-900 in Fig. 1(i) affirm the uniform dispersion of C, N, Pt, and Fe signals, which confirms that Pt and Fe are well-distributed throughout the entire matrix. To further elucidate the coordination environments of Pt and Fe SAs, extended X-ray absorption fine structure (EXAFS) analyses were carried out. As shown in Fig. 1(j), the Fe K-edge FT-EXAFS spectrum of PF-900 exhibits a pronounced peak at 1.5 Å corresponding to the Fe–N coordination shell, without any detectable Fe–Fe metallic contribution in the nearest-neighbor region. Likewise, the Pt K-edge FT-EXAFS spectrum of PF-900 in Fig. 1(k) reveals only a Pt–N peak at 1.6 Å, with no evident Pt–Pt contribution. These results confirm that both Fe and Pt species are atomically dispersed and coordinated predominantly with nitrogen.

Fig. 1 (a) Synthesis procedure of PF-900. (b) XRD pattern, (c) SEM image, (d) TEM image, and (e) HAADF-STEM image of PF-900. (f) HR HAADF-STEM images of PF-900, SAs and Pt–Fe hetero-pairs are indicated by the circles and rectangles, respectively. (g) Enlarged HAADF-STEM image in (f), and (h) intensity profile of the Pt–Fe hetero-pair (site #1), showing the different contrasts of Pt and Fe atoms. (i) HAADF-STEM image and corresponding EDS elemental mappings of C, N, Pt, and Fe signals of PF-900. (j) Fe K-edge and (k) Pt K-edge FT-EXAFS spectra of PF-900.

Ex situ pyrolysis of PF-ZIF under various temperatures

Based on our previous findings,⁵⁰ we confirm the presence of Fe SAs in both samples, where F-ZIF was pyrolyzed at 400 °C and 900 °C, denoted as F-400 and F-900, respectively. However, at a pyrolysis temperature of 1100 °C (denoted as F-1100), HAADF-STEM images reveal the complete absence of Fe SAs from the matrix of F-1100 [Fig. S4]. To investigate the role of Pt–Fe interaction in dual metal systems, we conducted a series of pyrolysis experiments at temperatures of 400 °C and 1100 °C denoted as PF-400 and PF-1100, respectively. These experiments were designed to highlight the difference compared to the single Fe system. As displayed in Fig. S5(a and b), TEM and HAADF-STEM images reveal the partial carbonization of matrix for PF-400. Despite only partial transformation, SAs were observed in the matrix, albeit at a low density [Fig. S5(c)]. Additionally, EDS elemental mappings [Fig. S6] confirm that Pt and Fe remain in the matrix of PF-400. The results for PF-400 are closely resemble those observed in F-400, where only a limited formation of SAs was detected at the relatively low pyrolysis temperature. However, after pyrolysis at 1100 °C, while the density of individual SAs decreased significantly, Pt–Fe hetero-pairs remained stable in PF-1100 [Fig. 2(a)]. In contrast, none of the SAs were preserved in F-1100. Moreover, the corresponding EDS elemental mappings in Fig. 2(a) and Fig. S4 reveal a notably lower Fe signal in F-1100 compared to PF-1100, indicating a substantial loss of Fe SAs in the absence of adjacent Pt–Fe pair interactions. To further support the conclusion, we performed a statistical analysis from multiple STEM images in Fig. S7. While the overall density of both isolated SAs and hetero-pairs decreased at 1100 °C relative to 900 °C, the relative proportion of hetero-pairs increased significantly, exceeding that of isolated SAs. This result exhibits that Pt–Fe hetero-pairs are more resistant to high-temperature pyrolysis and play a crucial role in stabilizing Fe atoms.

Fig. 2 HAADF-STEM images and EDS elemental mapping of (a) PF-1100, where SAs and Pt–Fe hetero-pairs are indicated by the circles and rectangles, respectively. (b) XRD patterns, (c) Raman spectra, (d) XPS Fe 2p, and (e) XPS Pt 4f spectra of PF-400, PF-900, and PF-1100.

Additionally, Fig. 2(b) presents the XRD patterns of the samples, where PF-400 retains the characteristic diffraction peaks of ZIF-8, while PF-900 and PF-1100 exhibit graphite-like diffraction features, confirming the phase transition during pyrolysis. Moreover, the Raman spectra in Fig. 2(c) reveal prominent D and G bands for PF-900 and PF-1100, with a lower I_D/I_G ratio in PF-1100, indicating a reduced nitrogen defect concentration due to enhanced graphitization. Fig. 2(d and e) display the XPS spectra of Fe 2p and Pt 4f, respectively. For PF-400, the Fe 2p spectrum indicates the presence of reduced Fe species, whereas in PF-900 the spectrum is dominated by a Fe³⁺ 2p_3/2 component at ∼710.8 eV, accompanied by characteristic satellite features, in agreement with the literature.^16,32,51 This confirms the absence of metallic Fe after pyrolysis at 900 °C. Upon further increasing the pyrolysis temperature to 1100 °C (PF-1100), the Fe 2p spectrum exhibits a slight positive shift while remaining consistent with Fe³⁺ species, suggesting a higher oxidation state compared to PF-900. Similarly, the Pt 4f spectrum of PF-1100 can be deconvoluted into Pt²⁺ and Pt⁴⁺ states, with 4f_7/2 peaks at ∼72.7 eV and ∼75.2 eV, respectively, accompanied by their corresponding 4f_5/2 components at +3.35 eV. These assignments are consistent with reported binding energies,^16,32,51 indicating modified Pt bonding coordination at elevated pyrolysis temperatures. Furthermore, the XPS N 1s spectra [Fig. S8] indicate that increasing the pyrolysis temperature to 900 °C leads to the emergence of multiple nitrogen configurations in Fe-900. Among these, pyridinic and pyrrolic nitrogen species are recognized as defect sites capable of effectively anchoring metal SAs. However, in PF-1100, the relative content of these nitrogen species, particularly pyridinic nitrogen, decreases significantly due to thermal degradation. This reduction is closely correlated with the diminished stability and density of single-atom configurations at higher temperatures. Combined with HAADF-STEM imaging results, these findings suggest that the formation of Pt–Fe hetero-pairs in the Pt–Fe dual-metal system stabilizes Fe atoms during high-temperature pyrolysis. The configuration of Pt–Fe hetero-pair modulates the electronic structure of both Fe and Pt, enhancing the electronic coupling between N-doped carbon matrix and metal atoms.^16,32 In addition to this strengthened metal–support interaction, long-range interactions between adjacent Pt and Fe atoms are likely to further modulate the local electronic environment, thereby suppressing aggregation and evaporation at elevated temperatures.^52,53 Moreover, DFT studies from the literature have demonstrated that the incorporation of Pt into Fe–N–C frameworks increases the cohesive energy between Fe atoms and the N-doped carbon matrix,⁵⁴ which is consistent with the stabilization effect observed in our system. Taken together, these factors highlight that the enhanced stability of Pt–Fe hetero-pairs arises from a synergistic combination of metal–metal interactions and reinforced coupling with the N-doped carbon support, effectively preventing Fe loss in contrast to single-metal systems.⁵⁵

In situ STEM observation during the pyrolysis of F-ZIF

To explore the stabilization of Pt–Fe hetero-pairs, we conducted additional in situ STEM experiments to observe the real-time structural evolution in F-ZIF and PF-ZIF under high temperatures. In our prior research, we examined the mechanism of SA formation in F-ZIF at pyrolysis temperatures below 400 °C, where Fe SAs remained evenly distributed within the matrix. However, this study revealed that, when exposed to an electron beam at temperatures above 500 °C, the Fe SAs acquired sufficient energy to migrate and aggregate into NPs. In situ STEM images in Fig. 3 display the structural evolution of atomic clustering during the pyrolysis of F-ZIF from 500 °C to 600 °C, the illustrations in the lower panels of the images depict the arrangements of the atoms in each corresponding state. At 500 °C, a few Fe atoms persisted as randomly dispersed SAs in the matrix, while others started to cluster, forming amorphous NCs [Fig. 3(a and b)]. As the temperature increased to 510 °C, the remaining SAs migrated and merged to form the NCs [Fig. 3(c)]. Between 510 °C and 530 °C, these activated NCs moved across the carbon matrix, attaching to and combining with other NCs [Fig. 3(c and d)]. Interestingly, at this stage, both merging and dissolution behaviors were observed among the NCs, with smaller NCs occasionally detaching from larger ones and existing independently, as indicated by the green arrows in the image [Fig. 3(d)]. Furthermore, Fe atoms gained additional energy, enabling them to detach from isolated NCs and migrate to neighboring ones at higher temperatures [Fig. 3(d and e)]. By the time the temperature reached 600 °C, most dispersed atoms had integrated into larger NCs, which displayed irregular shapes [Fig. 3(f)]. As the temperature was maintained at 600 °C for 140 s, the NCs, influenced by the Ostwald ripening effect, restructured into more regular shapes to minimize surface energy [Fig. 3(g)].

Fig. 3 (a) HAADF-STEM images of F-ZIF at 500 °C. (b)–(g) In situ STEM images and corresponding schematics revealing the evolutions of Fe atoms and NCs at the atomic scale from 500 °C to 600 °C for different time periods during pyrolysis.

Furthermore, in situ STEM images were recorded at temperatures ranging from 500 °C to 900 °C to examine the structural evolution [Fig. 4]. Fig. 4(a–e) presents STEM images captured at 500 °C, 600 °C, 700 °C, 800 °C, and 900 °C, respectively. As the temperature increased, the number of SAs in the matrix progressively decreased, while both the quantity and size of NCs increased, with some NCs further transforming into crystalline NPs. To quantitatively analyze the clustering behavior of Fe SAs, the projected area ratios of isolated SAs and NCs/NPs are shown in the lower panels of Fig. 4(a–e). At 500 °C, SAs accounted for approximately 57.4% of the projected area, while NCs/NPs comprised 42.6%. As the temperature rose, the proportion of SAs dropped sharply, and the share of NCs/NPs increased, suggesting that the merging of SAs into NCs/NPs prevailed over dissolution behavior at higher temperatures. Ultimately, the proportion of SAs reached a minimum of 8.6% at 500 °C, indicating that most SAs had merged into NCs/NPs. Additionally, the average diameter and density of NCs/NPs at various temperatures are depicted in Fig. 4f. The average diameter, initially 0.85 nm at 500 °C, exhibited a pronounced increase during pyrolysis, ultimately reaching 1.63 nm by 900 °C. Moreover, the standard deviation which reflects the size distribution, was narrow at 500 °C, indicating a well-defined NCs/NPs clustering at this stage. However, as the temperature increased, some of the NCs merged into larger entities, resulting in a broader size distribution, as evidenced by the increasing standard deviation. Meanwhile, the density of NCs/NPs exhibit slight fluctuations between 500 °C and 800 °C before experiencing a significant decline at 900 °C. This pattern implies that at lower temperatures (500 °C to 800 °C), only Fe SAs possess sufficient mobility to migrate and form smaller, well-dispersed NCs/NPs throughout the matrix. However, at higher temperatures of 900 °C, the energy threshold for NC migration and coalescence is met, leading to more significant merging and a consequent reduction in NC/NP density. Thus, NC/NP growth at lower temperatures is primarily driven by SA migration, whereas coalescence predominates at elevated temperatures. Furthermore, Fig. 4g shows a low-magnification STEM image of the sample post-pyrolysis, with yellow-dashed and blue-dashed squares marking regions of interest with and without electron beam exposure, respectively. In the beam-exposed region, HR HAADF-STEM images revealed that disordered NCs formed during pyrolysis transformed into crystalline NPs afterward [Fig. 4h]. Additionally, Fig. 4(j and k) display the HR HAADF-STEM image and corresponding FFT pattern of the NC in Fig. 4h, confirming the Fe₃C NP structure. The lattice spacings of 0.228 nm and 0.224 nm correspond to the (002) and (201) planes of Fe₃C, respectively [Fig. 4(j)]. Remarkably, in the region without beam exposure, Fe SAs remained fully intact in the matrix post-pyrolysis, showing no further clustering [Fig. 4i]. These findings suggest that the formation of Fe₃C NPs depends not only on temperature but also on electron beam irradiation. Moreover, the emergence of Fe₃C NPs indicates that Fe SAs is unstable above 500 °C under beam exposure. In control experiments, a higher concentration of Fe precursors was incorporated during ZIF synthesis (denoted as HF-ZIF) to investigate interactions between the carbon matrix and elevated Fe levels. Ex situ pyrolysis of HF-ZIF was performed at 900 °C (denoted as HF-900), revealing larger Fe₃C NPs in the matrix, consistent with the NC structures observed in F-ZIF under electron beam irradiation [Fig. S9]. This alignment underscores a robust interaction between Fe atoms and the carbon matrix.

Fig. 4 In situ STEM images of F-ZIF and corresponding projected area ratio between SAs and NCs/NPs at (a) 500 °C, (b) 600 °C, (c) 700 °C, (d) 800 °C, and (e) 900 °C. (f) The plot of average diameter and density of NCs/NPs at different temperatures. (g) HAADF-STEM image after 900 °C pyrolysis. (h) and (i) The magnified HAADF-STEM images of the sample w/ and w/o electron beam exposure, as indicated blue-square and yellow-square regions in (g), respectively. (j) The magnified HAADF-STEM images of the pink area in (h), the atomic structure of Fe₃C is presented in the image, orange and yellow spots denote Fe and C atoms, respectively. (k) corresponding FFT pattern showing the formation of Fe₃C structure.

In situ STEM observation during the pyrolysis of PF-ZIF

In addition to conducting the in situ pyrolysis experiments of F-ZIF, we also performed in situ STEM observation on PF-ZIF at elevated temperatures. Fig. 5a shows the low-magnification STEM image of PF-ZIF, where the high-contrast agglomerations with irregular shapes were found throughout the matrix at 200 °C [Fig. 5b, marked by the yellow arrows]. As the temperature increased to 300 °C, these species began to migrate randomly within the matrix. Then, they merged into sphere-like particles from 350 °C to 450 °C, disappearing entirely above 450 °C. These high-contrast particles are supposed to be the metal precursor, followed by decomposition and dispersion throughout the matrix at elevated temperatures.⁵⁶ Furthermore, the homogeneous structure remained unchanged throughout the rest of the pyrolysis process up to 900 °C [Fig. 5b and Fig. S10]. Remarkably, HR HAADF-STEM images reveal a high density of Pt–Fe hetero-pairs within the carbon matrix after pyrolysis, with no discernible NCs formed, as shown in Fig. 5(c). This contrasts sharply with observations of F-ZIF under electron beam exposure, where aggregation into NPs is evident in Fig. 4(h). The enhanced stability of Pt and Fe atoms under beam irradiation at elevated temperatures provides compelling evidence for strong Pt–Fe hetero-pairs interactions, likely governed by electronic structure modulation and metal–support coupling within the carbon matrix. In addition to Pt, recent studies have shown that other noble metals such as Rh,^15,57 as well as non-noble metals including Cu, Mn, Co, and Ni, can also form hetero-pairs with Fe and act as dual-atom catalysts.^52,58 These findings highlight not only a robust synthesis strategy for stable Pt–Fe DACs, but also the broader potential of hetero-pair design principles for advancing diverse dual-metal catalyst systems.

Fig. 5 (a) HAADF-STEM images of PF-ZIF at RT. (b) In situ sequential STEM images revealing the dispersion of metal precursors in PF-ZIF at various temperatures. (c) HR HAADF-STEM images of PF-ZIF after pyrolysis at 900 °C.

To explore the influence of metal precursor concentration on the stability and structural evolution of Pt–Fe system, we increased the concentration of Pt and Fe precursor loading in ZIF (denoted as HPF-ZIF) and conducted in situ STEM experiments, as shown in Fig. 6. In the early pyrolysis stage at 300 °C, in situ HR HAADF-STEM images in Fig. 6(a) reveals the coexistence of SAs, NCs, and larger clusters, indicating initial aggregation due to high metal concentration and surface energy. As temperature increased, Fig. 6(a–d) demonstrates NCs and clusters merging, followed by transformation into crystalline NPs at elevated temperatures, driven by thermal energy and electron beam-induced dynamics. A schematic of this structural evolution, illustrated in Fig. 6(e), depicts the sequence: dispersed SAs form initially, aggregate into NCs as temperature rises, and further coalesce into stable clusters, which crystallize into NPs via atomic rearrangement, confirmed by increasing diffraction spots in FFT patterns [lower-right panel in Fig. 6(a–d)]. Post-pyrolysis HR HAADF-STEM images [Fig. 6(f and g)] and Fig. S11 show that SAs completely transform into uniformly dispersed NPs across both beam-exposed and non-exposed regions, highlighting thermodynamic instability of SAs at high precursor concentrations. HR HAADF-STEM and FFT analysis in Fig. 6(h and i) identify these NPs as Pt-doped α-Fe phases, with an ordered atomic structure depicted in Fig. 6(j). Importantly, while high precursor loading suppresses single-atom stabilization, the preferential formation of Pt-doped α-Fe, rather than Fe or Pt carbides, strongly suggests a favorable Pt–Fe interaction. In this competition, Fe preferentially interacts with Pt, which weakens Fe–C bonding and suppresses the carbide pathway.^54,59–61 Consistent with the Fe–Pt phase diagram and the solid-solution behavior at low Pt concentrations, the system therefore evolves preferentially into Pt-doped α-Fe phases rather than Fe carbides. This observation underscores the critical role of metal–metal coupling in governing phase evolution and offers valuable mechanistic insight into the structural transformation pathways of dual-metal systems under high metal loading and harsh pyrolysis conditions.

Fig. 6 In situ STEM images and corresponding FFT patterns demonstrating the structural evolutions from SAs into crystalline Pt-doped α-Fe NPs at (a) 300 °C, (b) 400 °C, (c) 450 °C, and (d) 565 °C of HPF-ZIF. (e) Schematics of the evolutions from SAs to Pt-doped α-Fe NPs. (f)–(g) Low-mag and high-mag HAADF-STEM images of PF-ZIF after pyrolysis. (h)–(i) HR STEM image and the corresponding FFT pattern of the particle highlighted in (g), respectively. (j) Atomic structure of Pt-doped α-Fe NP.

Conclusions

In this study, we reveal the stabilization and structural evolution of Pt–Fe dual-metal SAs encapsulated within ZIF-derived N-doped carbon during high-temperature pyrolysis, offering transformative insights for advanced catalysts. Systematic in situ STEM and post-pyrolysis characterization studies reveal the evolution pathways of F-ZIF, highlighting electron beam irradiation as an additional energy source that induces Fe SA aggregation at 500 °C in F-ZIF and ultimately leads to the formation of Fe₃C NPs at 900 °C. In contrast, PF-ZIF sustains stable Pt–Fe hetero-pairs up to 900 °C, demonstrating superior resistance to aggregation and evaporation, driven by enhanced electronic coupling between metal atoms and N-doped carbon matrix. Increasing Pt and Fe precursor concentrations in HPF-ZIF further elucidates structural dynamics, with in situ STEM revealing a transition from SAs to amorphous clusters, then crystalline Pt-doped α-Fe NPs, reflecting robust Pt–Fe bonding and temperature-driven phase transitions. These findings, supported by high-resolution HAADF-STEM, establish a mechanistic framework for designing stable dual-metal materials. Our work paves the way for scalable synthesis strategies, with future studies focusing on operando stability, low-temperature methods, and heterometallic systems to optimize catalytic efficiency for industrial applications.

Author contributions

Conceptualization: Kai-Yuan Hsiao, Ming-Yen Lu. Methodology: Kai-Yuan Hsiao, Ming-Yen Lu. Investigation: Kai-Yuan Hsiao, Yi-Dong Lin, Yu-Ru Lin, Ching-Wei Chin, Chun-Hui Lin, Dun-Jie Jhan, Ruei-Hong Cyu, Yan-Gu Lin, Yu-Lun Chueh, and Ming-Yen Lu. Draft writing: Kai-Yuan Hsiao. Draft review and editing: Ming-Yen Lu.

Conflicts of interest

The authors declare no competing financial interests.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d5mh01324k.

Acknowledgements

This study was supported financially by the National Science and Technology Council (NSTC 112-2221-E-007-043, NSTC 113-2124-M-007-002 and NSTC 113-2221-E-007-052-MY3). We thank Mr Y. S. Chen of the Instrumentation Center at National Tsing Hua University for technical support.

Notes and references

1. X. Li, X. Yang, Y. Huang, T. Zhang and B. Liu, Adv. Mater., 2019, 31, 1902031.
2. A. Wang, J. Li and T. Zhang, Nat. Rev. Chem., 2018, 2, 65–81.
3. R. Chen, S. Chen, L. Wang and D. Wang, Adv. Mater., 2024, 36, 2304713.
4. Q. Yang, Y. Jiang, H. Zhuo, E. M. Mitchell and Q. Yu, Nano Energy, 2023, 111, 108404.
5. N. J. O’Connor, A. Jonayat, M. J. Janik and T. P. Senftle, Nat. Catal., 2018, 1, 531–539.
6. H. Xu, Y. Zhao, Q. Wang, G. He and H. Chen, Coord. Chem. Rev., 2022, 451, 214261.
7. J. Wan, W. Chen, C. Jia, L. Zheng, J. Dong, X. Zheng, Y. Wang, W. Yan, C. Chen and Q. Peng, Adv. Mater., 2018, 30, 1705369.
8. L. Zeng, K. Cheng, F. Sun, Q. Fan, L. Li, Q. Zhang, Y. Wei, W. Zhou, J. Kang and Q. Zhang, Science, 2024, 383, 998–1004.
9. X. Tang, J. Ye, L. Guo, T. Pu, L. Cheng, X. M. Cao, Y. Guo, L. Wang, Y. Guo and W. Zhan, Adv. Mater., 2023, 35, 2208504.
10. K. Naveen, T. Mahvelati-Shamsabadi, P. Sharma, S.-H. Lee, S. H. Hur, W. M. Choi, T. J. Shin and J. S. Chung, Appl. Catal., B, 2023, 328, 122482.
11. C.-C. Cheng, T.-Y. Lin, Y.-C. Ting, S.-H. Lin, Y. Choi and S.-Y. Lu, Nano Energy, 2023, 112, 108450.
12. S. Liang, T. Zhang, Y. Zheng, T. Xue, Z. Wang, Q. Wang and H. He, Appl. Catal., B, 2023, 333, 122801.
13. S. Chen, J. Chen, Y. Li, S. Tan, X. Liao, T. Zhao, K. Zhang, E. Hu, F. Cheng and H. Wang, Adv. Funct. Mater., 2023, 33, 2300801.
14. S. Wei, A. Li, J.-C. Liu, Z. Li, W. Chen, Y. Gong, Q. Zhang, W.-C. Cheong, Y. Wang and L. Zheng, Nat. Nanotechnol., 2018, 13, 856–861.
15. Y. Zhou, E. Song, W. Chen, C. U. Segre, J. Zhou, Y. C. Lin, C. Zhu, R. Ma, P. Liu and S. Chu, Adv. Mater., 2020, 32, 2003484.
16. A. Han, X. Wang, K. Tang, Z. Zhang, C. Ye, K. Kong, H. Hu, L. Zheng, P. Jiang and C. Zhao, Angew. Chem., Int. Ed., 2021, 60, 19262–19271.
17. X. Li, Y. He, S. Cheng, B. Li, Y. Zeng, Z. Xie, Q. Meng, L. Ma, K. Kisslinger and X. Tong, Adv. Mater., 2021, 33, 2106371.
18. P. Balasubramanian, H. Khan, J. H. Baek and S. H. Kwon, Chem. Eng. J., 2023, 471, 144378.
19. T. Najam, S. S. A. Shah, S. Ibraheem, X. K. Cai, E. Hussain, S. Suleman, M. S. Javed and P. Tsiakaras, Energy Storage Mater., 2022, 45, 504–540.
20. L. J. Cao, X. L. Wang, C. Yang, J. J. Lu, X. Y. Shi, H. W. Zhu and H. P. Liang, ACS Sustainable Chem. Eng., 2021, 9, 189–196.
21. P. C. Deng, J. L. Duan, F. L. Liu, N. Yang, H. B. Ge, J. Gao, H. F. Qi, D. Feng, M. Yang, Y. Qin and Y. J. Ren, Angew. Chem., Int. Ed., 2023, 62, e202307853.
22. Z. C. Fan, H. Wan, H. Yu and J. J. Ge, Chin. J. Catal., 2023, 54, 56–87.
23. Y. Zhang, C. P. Jin, C. C. Wang, X. Zeng, M. Yang, C. J. Hou and D. Q. Huo, Biosens. Bioelectron., 2025, 271, 117001.
24. Q. H. Wang, H. J. Liang, J. W. Zhou, J. K. Wang, Z. Ye, M. Zhao, H. M. Yang, Y. H. Song and J. J. Guo, Chem. Eng. J., 2023, 467, 143482.
25. W. Y. Zhang, S. Y. Yi, Y. H. Yu, H. Liu, A. Kucernak, J. Wu and S. Li, J. Mater. Chem. A, 2023, 12, 87–112.
26. Y. F. Liu, R. Duan, X. Li, L. Luo, J. Gong, G. F. Zhang, Y. J. Li and Z. Li, Inorg. Chem., 2022, 61, 13210–13217.
27. S. Wang, Z. F. Hu, Q. L. Wei, P. Cui, H. M. Zhang, W. N. Tang, Y. Q. Sun, H. Q. Duan, Z. C. Dai, Q. Y. Liu and X. W. Zheng, ACS Appl. Mater. Interfaces, 2022, 14, 20669–20681.
28. W. Q. Ma, C. Y. Jing, P. Wu and W. Y. Li, Mol. Catal., 2024, 567, 114458.
29. W. Q. Liu, Q. J. Chen, F. Zhang, D. Y. Xu and X. J. Li, Int. J. Hydrogen Energy, 2021, 46, 13180–13189.
30. G. D. Cheng, F. Y. Chen, S. L. Li, Y. Hu, Z. C. Dai, Z. F. Hu, Z. B. Gan, Y. Q. Sun and X. W. Zheng, J. Mater. Chem. B, 2024, 12, 1512–1522.
31. W. M. Gong, P. Guo, L. Zhang, R. Fu, M. L. Yan, C. Wu, M. M. Sun, G. H. Su, Y. Y. Wang, J. S. Ye, H. B. Rao and Z. W. Lu, Chem. Eng. J., 2025, 507, 160400.
32. W.-S. Song, M. Wang, X. Zhan, Y.-J. Wang, D.-X. Cao, X.-M. Song, Z.-A. Nan, L. Zhang and F. R. Fan, Chem. Sci., 2023, 14, 3277–3285.
33. J. Y. Wan, H. Zhang, J. Yang, J. G. Zheng, Z. K. Han, W. T. Yuan, B. R. Lan and X. D. Li, Appl. Catal., B, 2024, 347, 123816.
34. J. R. Yang, D. Q. Zeng, J. Li, L. Q. Dong, W. J. Ong and Y. L. He, Chem. Eng. J., 2021, 404, 126376.
35. B. F. Zhang, X. Q. Li, P. A. Bingham, K. Akiyama and S. Kubuki, Chem. Eng. J., 2023, 451, 138574.
36. R. Liu, R. Tang, J. Feng and T. Meng, Chem. Eng. J., 2023, 470, 144261.
37. R. Li, C. Gu, P. Rao, P. Deng, D. Wu, J. Luo, J. Li, Z. Miao, C.-W. Zheng and C. Shen, Chem. Eng. J., 2023, 468, 143641.
38. X. W. Zhong, S. L. Ye, J. Tang, Y. M. Zhu, D. J. Wu, M. Gu, H. Pan and B. M. Xu, Appl. Catal., B, 2021, 286, 119891.
39. P.-W. Hsieh, C.-C. Chi, C.-M. Wu, K.-Y. Hsiao and M.-Y. Lu, Appl. Surf. Sci., 2024, 670, 160612.
40. P.-H. Feng, K.-Y. Hsiao, D.-J. Jhan, Y.-L. Chen, P. Y. Keng, S.-Y. Chang and M.-Y. Lu, ACS Appl. Mater. Interfaces, 2023, 15, 48543–48550.
41. L. Zhang, Y. Li, L. Zhang, K. Wang, Y. Li, L. Wang, X. Zhang, F. Yang and Z. Zheng, Adv. Sci., 2022, 9, 2200592.
42. K.-Y. Hsiao, Y.-H. Tseng, C.-L. Chiang, Y.-D. Chen, Y.-G. Lin and M.-Y. Lu, ACS Appl. Nano Mater., 2023, 6, 10713–10724.
43. S. Dai, Y. You, S. Zhang, W. Cai, M. Xu, L. Xie, R. Wu, G. W. Graham and X. Pan, Nat. Commun., 2017, 8, 204.
44. C. Zhang, W. Sandorf and Z. Peng, ACS Catal., 2015, 5, 2296–2300.
45. L. DeRita, J. Resasco, S. Dai, A. Boubnov, H. V. Thang, A. S. Hoffman, I. Ro, G. W. Graham, S. R. Bare and G. Pacchioni, Nat. Mater., 2019, 18, 746–751.
46. G. Wang, M. Zhang, G. Zhang, Z. Wang, X. Chen, X. Ke, C. Wang, S. Chu and M. Sui, Adv. Funct. Mater., 2024, 34, 2308876.
47. K. Liu, Z. Sun, X. Peng, X. Liu, X. Zhang, B. Zhou, K. Yu, Z. Chen, Q. Zhou and F. Zhang, Nat. Commun., 2025, 16, 2167.
48. Y.-H. Tseng, K.-Y. Hsiao, C.-C. Chi and M.-Y. Lu, Mater. Today Energy, 2023, 38, 101425.
49. S.-Y. Su, Y.-T. Fan, Y.-J. Su, C.-W. Huang, M.-H. Tsai and M.-Y. Lu, J. Alloys Compd., 2021, 851, 156909.
50. K.-Y. Hsiao, Y.-D. Lin, Y.-R. Lin, C.-W. Chin, C.-H. Lin, R.-H. Cyu, Y.-G. Lin, Y.-L. Chueh and M.-Y. Lu, J. Mater. Chem. A, 2025 10.1039/d4ta07390h.
51. S. Wang, Z. Hu, Q. Wei, P. Cui, H. Zhang, W. Tang, Y. Sun, H. Duan, Z. Dai and Q. Liu, ACS Appl. Mater. Interfaces, 2022, 14, 20669–20681.
52. W. Liu, Z. Liang, S. Jing, J. Zhong, N. Liu, B. Liao, Z. Song, Y. Huang, B. Yan and L. Gan, Angew. Chem., 2025, 137, e202503493.
53. M. Tamtaji, Q. Peng, T. Liu, X. Zhao, Z. Xu, P. R. Galligan, M. D. Hossain, Z. Liu, H. Wong and H. Liu, Nano Energy, 2023, 108, 108218.
54. G. Bae, M. M. Kim, M. H. Han, J. Cho, D. H. Kim, M.-T. Sougrati, J. Kim, K.-S. Lee, S. H. Joo and W. A. Goddard III, Nat. Catal., 2023, 6, 1140–1150.
55. Q. Ma, H. Jin, J. Zhu, Z. Li, H. Xu, B. Liu, Z. Zhang, J. Ma and S. Mu, Adv. Sci., 2021, 8, 2102209.
56. B. Pal and M. Sharon, Thin Solid Films, 2000, 379, 83–88.
57. T. Liu, Y. Chen, A. Xu, X. Liu, D. Liu, S. Li, H. Huang, L. Xu, S. Jiang and Q. Luo, Sci. China: Chem., 2024, 67, 1352–1359.
58. Q. Wang, B. Chu, C. Shang, B. Shao, F. Yang, D. Dang, L. Li, M. Gu, X. Xiao and Q. Xu, Adv. Mater., 2025, e14343.
59. A. K. Mechler, N. R. Sahraie, V. Armel, A. Zitolo, M. T. Sougrati, J. N. Schwämmlein, D. J. Jones and F. Jaouen, J. Electrochem. Soc., 2018, 165, F1084.
60. F. Xiao, Q. Wang, G.-L. Xu, X. Qin, I. Hwang, C.-J. Sun, M. Liu, W. Hua, H.-W. Wu and S. Zhu, Nat. Catal., 2022, 5, 503–512.
61. N. A. Ishiki, K. Teixeira Santos, N. Bibent, K. Kumar, I. Reichmann, Y.-P. Ku, T. Asset, L. Dubau, M. Mermoux and H. Ge, Nat. Commun., 2025, 16, 6404.

---