Atomic-scale surface design for tailored nucleation in stable multivalent metal anodes

Abstract

Achieving uniform and reversible magnesium (Mg) deposition is a critical bottleneck for the practical implementation of Mg metal batteries (MMBs), as uncontrolled nucleation and dendritic growth undermine interfacial stability and cycling performance. To address this, we introduce an atomic-level surface design strategy that guides Mg nucleation through precise interface engineering. To model this concept, we designed a freestanding porous carbon nanofiber framework embedded with Zn single atoms (ZnSA@PCF), derived from pyrolyzed electrospun PAN/ZIF-8 composites. This architecture simultaneously provides high surface area via uniformly distributed hollow nanocages and magnesiophilic Zn single-atom sites that serve as catalytic centers to direct Mg plating. This dual design significantly reduces the nucleation overpotential and enables dendrite-free Mg growth up to 5 mA h cm⁻². The theoretical simulation results reveal strong Mg affinity at the introduced Zn SAC sites, while electrochemical tests demonstrate a high critical current density (17 mA cm⁻²) and ultra-stable cycling over 1500 h with 99.79% coulombic efficiency. This work establishes atomic-level catalyst engineering as a compelling paradigm for interfacial control in next-generation reversible MMBs.

---

Introduction

MMBs have gained significant attention for their numerous advantages, including their high theoretical volumetric capacity (3833 mA h cm⁻³),¹ the high natural abundance of magnesium (Mg, the 7th most naturally abundant element in Earth's crust),^2,3 and low material cost (Mg: $0.26 g⁻¹vs. Li: $1.44 g⁻¹).^4,5 Also, Mg metal has a lower tendency to form dendrites, giving rise to its reputation as one of the safest anode materials among metal batteries.^6–8 The intrinsically low self-diffusion energy barrier of Mg metal enables relatively packed plating,^9,10 while the strong metallic bonding between Mg atoms favors the formation of three-dimensional hemispherical deposits.¹¹ These theoretical findings support the view that MMBs are inherently safer against thermal runaway caused by short circuits. Despite this inherent advantage, these findings are primarily based on the intrinsic material properties of Mg and do not fully account for the complexities of the electrochemical environment. Recent studies have demonstrated that under practical cell conditions, including elevated current densities and extended plating durations, Mg dendrites can grow.^12–15 Therefore, achieving uniform and reversible Mg plating and stripping is essential for the practical advancement of MMBs.

To address these challenges, numerous strategies have been proposed, including electrolyte modification,^16–19 the use of functional additives,^20–22 artificial interphase construction,^23–26 and current collector engineering.^27–36 Among these, the modification of current collectors into three-dimensional (3D) architectures has been widely applied, as their porous structures reduce the local nucleation current density and effectively help accommodate volume changes during cycling.^{28,32,37–39} Additionally, the introduction of Mg-philic functional groups into the host structure can facilitate uniform nucleation, thereby enabling reversible Mg plating/stripping behavior. In particular, several Mg-philic catalysts (e.g., Au, Ag, Ni) have been introduced to promote uniform nucleation.^{12,27,30–33} Nonetheless, these conventional catalysts suffer from limitations such as poor spatial controllability and detachment during repeated cycling, which hinder long-term stability. In this context, the single-atom catalyst (SAC) strategy has emerged as a compelling alternative in metal battery systems, including lithium, sodium, and zinc.^40–46 SACs offer maximized atomic utilization and tunable coordination environments, which enhance electrochemical activity and stability at the atomic scale. These characteristics have shown great promise in facilitating uniform metal nucleation and suppressing dendrite growth. However, to date, the application of SACs in MMBs remains largely unexplored, despite their potential to address key interfacial challenges.

In this study, we present a rationally designed strategy to guide uniform and reversible Mg deposition via atomic-level catalyst engineering. Specifically, we developed a freestanding electrode composed of porous carbon nanofibers embedded with Zn single atoms (ZnSA@PCF), fabricated by controlling the pyrolysis of electrospun PAN/ZIF-8 precursors. This design achieves dual functionality: (i) the formation of abundant hollow carbon nanocages (∼50 nm) uniformly distributed throughout the fiber matrix provides a large surface area and abundant Mg²⁺ adsorption sites; and (ii) atomically dispersed Zn sites anchored within the N-doped carbon matrix act as magnesiophilic catalytic centers, effectively reducing the nucleation overpotential and promoting homogeneous Mg seed formation. Spectroscopic analyses confirm that Zn exists as Zn–N₃ single-atom sites, while density functional theory (DFT) calculations reveal significantly enhanced Mg adsorption energy. As a result, ZnSA@PCF enables dendrite-free Mg deposition up to 5 mA h cm⁻², exhibits a high critical current density (17 mA cm⁻²), and demonstrates ultra-stable cycling over 1500 h. Moreover, full-cell tests with a Mo₆S₈ cathode validate its practical applicability, achieving long-term cycling and excellent rate capability. These findings establish ZnSA@PCF as a promising scaffold for next-generation MMBs by simultaneously addressing the challenges of interfacial instability and uncontrolled nucleation.

Results and discussion

To address the persistent challenge of non-uniform Mg deposition, we rationally designed a porous carbon nanofiber host embedded with atomically dispersed Zn single atoms (ZnSA@PCF). This design fulfills two key criteria: (i) the formation of numerous hollow carbon cages along the nanofiber strands provides abundant nucleation sites and facilitates uniform Mg seeding, and (ii) the atomically dispersed Zn sites serve as highly magnesiophilic centers to promote homogeneous nucleation. To realize this architecture, ZnSA@PCF was fabricated via electrospinning a precursor solution composed of polyacrylonitrile (PAN) and a Zn-based zeolitic imidazolate framework (ZIF-8), followed by thermal treatment, as illustrated in Fig. 1a. The process began with the synthesis of polyhedral ZIF-8 nanoparticles via a simple precipitation method, yielding uniform particles with an average diameter of ∼50 nm (Fig. S1a) and no detectable byproducts (Fig. S1b). Subsequently, PAN and ZIF-8 were mixed into a solution, which was electrospun to produce uniform ZIF-8/PAN nanofibers. These nanofibers were then heat treated at 800 °C to finally produce ZnSA@PCF. During the post-heating process, organic components such as PAN and 2-methylimidazole were pyrolyzed into an N-doped carbon matrix; simultaneously, Zn ions were reduced into metal nodes. Interestingly, the strong interaction between the PAN and ZIF-8 during pyrolysis induced outward shrinkage of ZIF-8, transforming the ZIF-8 nanoparticles into hollow-structured carbon polyhedrons.^47,48 This resulted in the formation of numerous meso/macropores of uniform size in the resultant nanofibers after the heating process.

Fig. 1 Structural design and morphological characterization of fiber-based freestanding anode substrates. (a) Schematic of ZnSA@PCF synthesis. (b) Conceptual diagram for the freestanding anode substrates (CF, PCF, and ZnSA@PCF). SEM images of (c) CF, (d) PCF, and (e) ZnSA@PCF. Insets show cross-sectional views of individual fiber strands.

In addition to ZnSA@PCF, two control samples were prepared, as illustrated in Fig. 1b. One consisted solely of PAN-derived carbon fiber (CF), while the other, referred to as porous carbon nanofiber (PCF), was synthesized by co-electrospinning PAN with ZIF-8, followed by heat treatment at 950 °C. Zn is known to evaporate at temperatures above 907 °C,⁴⁹ enabling its complete removal during the high-temperature treatment. All substrates (CF, PCF, and ZnSA@PCF) were fabricated as self-standing electrodes with a diameter of ∼12 mm (Fig. S2).

We examined the morphology of the prepared fibers using scanning electron microscopy (SEM). The CF derived from the PAN nanofibers without ZIF-8 exhibited smooth surfaces (Fig. 1c) and lacked any visible pores along the fiber strands (inset image). By contrast, the PCF (Fig. 1d) and ZnSA@PCF (Fig. 1e) nanofibers exhibited rough surfaces owing to the protrusion of the ZIF-8 nanoparticles and the abundant meso/macropores on their surfaces. The overall morphology of each host was essentially unchanged compared with that before heat treatment (Fig. S3). Notably, interconnected conductive networks, featuring abundant meso/macropores both on the surface and within the nanofiber, provide efficient electronic pathways that help minimize local current density.

To more clearly investigate the unique pore structure of the ZnSA@PCF, transmission electron microscopy (TEM) analysis was conducted (Fig. 2a). The TEM image reveals hollow carbon nanocages (∼50 nm in diameter), derived from ZIF-8 nanoparticles, uniformly distributed throughout the nanofibers. To analyze the pore structure of the ZnSA@PCF in more depth, we measured its N₂ adsorption/desorption (Fig. 2b and c). According to the International Union of Pure and Applied Chemistry classification, ZnSA@PCF displays type I/IV isotherms with H2-type hysteresis, indicating a hierarchical porous structure with micropores and meso/microcavities. The steep desorption branch (P/P₀ = 0.5) indicates cavitation-driven emptying of the meso/macropores. Conversely, the CFs exhibit a reversible type-II isotherm of nonporous materials with a low specific surface area (29.6 m² g⁻¹), while ZnSA@PCF exhibits a significantly higher value (363.5 m² g⁻¹). PCF shows a similar isotherm with an even higher surface area (538.0 m² g⁻¹), likely owing to enhanced micropore formation resulting from further Zn removal during high carbonization temperatures. Such a hierarchical porosity enhances mechanical flexibility by dispersing mechanical stress. As shown in Fig. S4, both ZnSA@PCF and PCF recover their original shapes after bending, whereas the CFs do not. Therefore, the mechanical robustness of ZnSA@PCF is conducive to maintaining structural integrity during cell assembly and cycling.

Fig. 2 Structural and compositional characterization of ZnSA@PCF. (a) High-resolution TEM image of ZnSA@PCF. (b) N₂ adsorption–desorption isotherms and (c) Barrett–Joyner–Halenda pore size distribution of ZnSA@PCF. (d) SAED pattern and (e) EDS elemental mapping of C, N, and Zn in ZnSA@PCF. (f) High-resolution Zn 2p XPS spectrum of ZnSA@PCF, and N 1s spectra of (g) PCF and (h) ZnSA@PCF. (i) Comparison of the relative atomic ratios of nitrogen species in ZnSA@PCF and PCF.

Clarifying whether Zn is dispersed as metallic seeds or at the atomic scale is important. The selected area electron diffraction (SAED) pattern of ZnSA@PCF (Fig. 2d) shows no distinct crystalline lattices, confirming the absence of metallic Zn particles. Furthermore, metallic Zn was not detected in the X-ray diffraction (XRD) pattern (Fig. S5). Meanwhile, elemental mapping analysis (Fig. 2e) verifies that Zn is uniformly distributed, indicating that Zn exists in an atomic-scale dispersion without aggregation. Additionally, the TEM images and SAED patterns for PCF and CF are provided in Fig. S6. The atomic-scale dispersion of Zn is closely associated with its volatility. During pyrolysis, Zn atoms tend to aggregate into nanoparticles owing to their high surface energy, and these are easily volatilized at high temperatures. By contrast, non-aggregated Zn remains embedded in the carbon matrix, forming atomically dispersed Zn single-atom sites (ZnSA).

We further evaluated the chemical state of ZnSA@PCF using high-resolution Zn 2p X-ray photoelectron spectroscopy (XPS) in Fig. 2f and S7. The spin–orbit splitting (23.0 eV) between the two characteristic peaks (Zn 2p_3/2 and Zn 2p_1/2) indicates that Zn exists predominantly in the Zn²⁺ oxidation state.⁵⁰ In addition, we analyzed the high-resolution N 1s XPS spectra of both ZnSA@PCF and PCF to investigate the coordination environment. Both spectra can be deconvoluted into four distinct peaks assigned to pyridinic N (398.2 eV), Zn–N (399.4 eV), pyrrolic N (400.6 eV), and graphitic N (401.8 eV). ZnSA@PCF (Fig. 2g) exhibits a distinct Zn–N peak, indicating the successful coordination of Zn to nitrogen sites. On the other hand, PCF (Fig. 2h) shows a noticeable increase in graphitic N and a decrease in Zn–N. As summarized in Fig. 2i, the proportional changes of N 1s components in ZnSA@PCF and PCF clearly demonstrate this trend. The proportion of Zn–N in PCF is significantly lower than that in ZnSA@PCF, which is attributed to the evaporation of a substantial amount of Zn species during the high-temperature carbonization. This variation in Zn content was confirmed using inductively coupled plasma optical emission spectroscopy, revealing a significantly lower Zn loading in PCF (1.86 wt%) than that in ZnSA@PCF (9.93 wt%) (Table S1). Notably, the Zn loading in ZnSA@PCF is substantially higher than that reported in previous studies on Zn SACs, which typically range from 1 to 3 wt%.^51–54 This result highlights the effectiveness of our synthetic strategy in achieving a high loading of atomically dispersed Zn sites without aggregation.

To directly confirm that Zn is atomically dispersed, X-ray absorption spectroscopy (XAS) was employed, as it provides the oxidation state and coordination environment of Zn. Fig. 3a presents the Zn K-edge X-ray absorption near-edge structure (XANES) spectra of ZnSA@PCF, along with those of standard Zn foil, ZnO, and Zn(II) phthalocyanine (ZnPc). The absorption edge of ZnSA@PCF is located between those of the Zn foil and ZnO, indicating that the Zn atoms are positively charged and coordinated with N atoms. In addition, the oxidation number of ZnSA@PCF (Fig. 3b and S8) was determined to be 1.36, lower than that of ZnPc (2.0), which features a Zn–N₄ configuration. The calculation methodology is described in Fig. S9. Such a difference in the oxidation state implies that ZnSA in ZnSA@PCF have different coordination environments and electronic structures than those in ZnPc.

Fig. 3 Atomic structure analysis and theoretical investigation of Zn single-atom sites in ZnSA@PCF. (a) Zn K-edge XANES spectra and (b) oxidation state of Zn calculated from XANES fitting. (c) Extended EXAFS spectra of ZnSA@PCF and reference samples (Zn foil, ZnO and ZnPc). Wavelet transform (WT) analysis for the k²-weighted EXAFS signals for (d) Zn foil, (e) ZnO, and (f) ZnSA@PCF. (g) Geometric model of Zn–N–C coordination environment. (h) DFT-calculated adsorption energies and optimized structures of Mg adsorbed on ZnSA@PCF. (i) Formation energies of Zn single atoms coordinated with various N and C doping configurations.

Fig. 3c depicts the Zn K-edge Fourier-transform extended X-ray absorption fine structure (FT-EXAFS) spectra of the samples. In the ZnSA@PCF spectrum, a weak peak corresponding to the Zn–N first coordination shell is evident at 1.53 Å. Additionally, the Zn–Zn peaks corresponding to Zn foil (2.27 Å) or ZnO (2.91 Å) were not observed in the ZnSA@PCF spectrum. These results directly demonstrate the atomically monodispersed character of Zn and confirm the formation of single-atom Zn. Notably, the Zn–N peak intensity of the ZnSA@PCF is lower than that of ZnPc, implying a decrease in the coordination number of Zn. According to the fitting EXAFS curves in Fig. S9 and Table S2, the calculated average coordination number of the Zn atoms is ∼3.0, indicating that each Zn atom is coordinated to three N atoms, forming Zn–N₃ sites. The formation of this Zn–N₃ coordination structure is primarily attributed to high-temperature pyrolysis, which promotes Zn–N bond cleavage and transforms the conventional Zn–N₄ configuration into a lower coordination state. The monodispersed nature of the Zn atoms is supported by wavelet transform (WT) EXAFS analysis. Compared with the contour plots of the standard Zn foil (Fig. 3d) and ZnO (Fig. 3e), that of ZnSA@PCF (Fig. 3f) contains only one WT maximum at 5.1 Å⁻¹. This peak can be attributed to the Zn–N bond, further evidencing that ZnSA@PCF comprises atomically dispersed Zn sites.

To fundamentally understand how catalytic ZnSA influences Mg nucleation, DFT calculations were performed. Because strong metal ion adsorption significantly affects initial nucleation behavior,^55,56 we calculated and compared the Mg adsorption energy (E_ads) at different sites. First, we defined the various nucleation sites (geometric model of Zn–N–C coordination structure), such as pure graphene (C), N_x–C (N₂–C, N₃–C), and Zn–N_xC_y (Zn–N₂C₂, Zn–N₃C₁) (Fig. 3g). Among the geometric models with the same coordination number but different configurations, the structure with the lowest formation energy was selected (Fig. S10). We then calculated the adsorption energy of the Mg atom in each geometry (Fig. 3h). Compared with pure graphene (−0.253 eV), E_ads was slightly increased by N doping (−0.269 eV), regardless of the number of N atoms. Notably, the introduction of Zn significantly enhances the Mg adsorption, and E_ads was increased by the coordination environment of Zn (−0.430 eV on Zn–N₂C₂ and −0.690 eV on Zn–N₃C₁). Additionally, all possible adsorption sites with E_ads are provided in Fig. S11 and S12. To fully understand the origin of the strong adsorption in Zn–N–C, we performed crystal orbital Hamilton population (COHP) analysis between Mg and the adsorption site(s) of surface atom(s) (Fig. S13). These results support that Mg more strongly interacts with the N-coordinated Zn site.

To elucidate the role of nitrogen in the ZnSA@PCF system, we examined the formation energy (E_form) of different Zn coordination environments. Nitrogen coordination is widely recognized for improving the stability of SACs on carbon supports.⁵⁷ Specifically, we compared Zn–C₄, Zn–N₂C₂, and Zn–N₃C₁ configurations, with calculated E_form values of 4.969, 2.781, and 1.671 eV, respectively (Fig. 3i). The progressive decrease in E_form clearly indicates that N atoms significantly enhance the thermodynamic stability of Zn sites. Although this coordination does not directly improve Mg adsorption, it is essential for securely anchoring the Zn atoms within the carbon matrix, thereby enabling their catalytic function.

To gain insights into the actual Mg plating behavior on ZnSA@PCF, the morphological evolution of Mg seeds was examined with increasing capacity. In the case of the Cu foil (Fig. 4a–c), Mg deposition resulted in numerous randomly distributed large aggregates, which were prone to delamination from the flat 2D surface. For the CF electrode (Fig. 4d–f), similar to Cu foil, Mg was found to deposit non-uniformly across the fiber surface, indicating poor nucleation control. By contrast, the PCF sample (Fig. 4g–i) initially facilitated a more uniform dispersion of Mg deposits. However, with extended deposition time, localized aggregation of Mg became apparent, indicating limited stability in nucleation uniformity. Notably, ZnSA@PCF exhibited highly uniform Mg deposition across the fiber surface (Fig. 4j–l), accompanied by the development of crystalline domains. The crystalline phase of metallic Mg was observed in the XRD pattern following Mg deposition (0.5 mA h cm⁻²) onto ZnSA@PCF (Fig. S14). These Zn single atoms were uniformly dispersed throughout the carbon host, facilitating uniform and efficient Mg deposition at high areal capacities of up to 5 mA h cm⁻² (Fig. S15).

Fig. 4 Morphological evolution of Mg deposits on various substrates during electroplating. SEM images of Mg deposits on various substrates (Cu, CF, PCF, and ZnSA@PCF); initial state, after 0.5 mA h cm⁻² deposition, and after 2 mA h cm⁻² deposition onto (a–c) Cu, (d–f) CF, (g–i) PCF, and (j–l) ZnSA@PCF, all at a current density of 0.5 mA cm⁻².

This improved deposition behavior is attributed to the synergistic effect of the high surface area and the well-dispersed ZnSA nucleation sites, which effectively guide Mg nucleation and growth. Notably, XAS analysis (Fig. S16) demonstrated that these Zn single atoms undergo negligible structural changes after cycling. This structural integrity confirms their sustained role as effective nucleation sites for Mg, even after extended cycling. In this architecture, Mg metal preferentially nucleates at single-atom sites with a strong adsorption energy, promoting homogeneous Mg nucleation and effectively suppressing Mg dendrite formation. Overall, the ZnSA@PCF structure enables homogeneous Mg plating, thereby extending battery lifespan through improved interfacial stability.

We next examined how the guided uniform Mg nucleation influences the electrochemical performance. Asymmetric cells were assembled using Mg metal as the counter electrode, while CF, PCF, and ZnSA@PCF served as working electrodes. As shown in Fig. 5a, the nucleation overpotential at a current density of 5 mA cm⁻² was markedly lower for ZnSA@PCF (423 mV) than for PCF (645 mV) and CF (648 mV). This trend persisted across various current densities (Fig. 5b and S17), with ZnSA@PCF consistently exhibiting the lowest nucleation overpotential. The lower overpotential of PCF than that of CF is primarily attributed to its larger specific surface area, which helps alleviate local current density. Notably, despite ZnSA@PCF and PCF having comparable surface areas, the overpotential of ZnSA@PCF was further decreased. This result clearly confirms that, beyond the surface area effect, the presence of Zn SAC provides additional active nucleation sites, facilitating Mg nucleation.

Fig. 5 Electrochemical performance of ZnSA@PCF as a current collector for MMBs. (a) Nucleation overpotential of Mg plating on various substrates (CF, PCF, and ZnSA@PCF) at a current density of 0.5 mA cm⁻². (b) Comparison of nucleation overpotentials at various current densities. (c) Onset reduction potentials from CV curves at a scan rate of 1 mV s⁻¹. (d) Galvanostatic cycling performance with CE test at 0.5 mA cm⁻². (e) Comparison of electrochemical performance of ZnSA@PCF with previously reported works. (f) Cycling performance of full cells composed of a Mo₆S₈ cathode and various current collectors as the anode substrate at (1C = 128 mA h cm⁻²). (g) Rate capability of Mo₆S₈ full cells.

Linear sweep voltammetry (LSV) performed at a scan rate of 1 mV s⁻¹ further provides insight into the different Mg plating kinetics (Fig. 5c). Among the hosts, ZnSA@PCF exhibited the lowest onset reduction potential at −0.16 V, followed by PCF (−0.19 V) and CF (−0.23 V), indicating a sequential enhancement in Mg reduction kinetics. These findings are in good agreement with the electrochemical impedance spectroscopy test results (Fig. S18), where ZnSA@PCF exhibited the lowest charge transfer resistance.

To verify the stable plating/stripping behavior of ZnSA@PCF, its long-term cycling performance was investigated. In Fig. 5d and S19, ZnSA@PCF maintained a stable average overpotential of 50 mV for over 1500 h, with a high coulombic efficiency (99.79%). This remarkable electrochemical stability is attributed to the homogeneous distribution of Zn SACs, which promote uniform and reversible Mg plating/stripping, effectively mitigating the formation of dendritic or localized deposits. By contrast, both CF and PCF suffered from early cell failure under high current densities, primarily owing to the absence of stable nucleation sites that led to inhomogeneous Mg deposition. Notably, compared with the reported analogs, ZnSA@PCF exhibited best-in-class plating/stripping characteristics (Fig. 5e and Table S3). The impact of zinc SACs on Mg deposition/stripping kinetics was further investigated under various current densities (Fig. S20). ZnSA@PCF exhibited stable Mg deposition/stripping profiles with relatively low overpotentials across all applied current densities. Impressively, even after applying a high current density of 15 mA cm⁻², it returned to 1 mA cm⁻² with minimal voltage fluctuation, highlighting the excellent reversibility of ZnSA@PCF, even under harsh electrochemical conditions. To determine the threshold for short-circuit formation, a critical current density (CCD) test was conducted (Fig. S21). ZnSA@PCF achieved a high CCD of 17 mA cm⁻², exhibiting a consistent trend with its excellent rate performance.

We further assessed the practical applicability of ZnSA@PCF in full-cell configurations. A Chevrel phase (Mo₆S₈) cathode was employed, with synthesis details described in the Experimental Section. Its morphological and crystalline characteristics were also characterized (Fig. S22). At a current rate of 1C (128 mA h g⁻¹), the ZnSA@PCF‖Mo₆S₈ full cell maintains a stable average discharge capacity of 72.5 mA h g⁻¹ over 1200 cycles, accompanied by nearly 100% coulombic efficiency (Fig. 5f). By contrast, full cells assembled with CF and PCF anodes failed after approximately 600 and 800 cycles, respectively. Moreover, the ZnSA@PCF-based full cell demonstrated outstanding rate performance over the current density range of 0.2C to 5C (Fig. 5g and S23). These findings highlight the superior cycling stability and rate capability of ZnSA@PCF, underscoring its strong potential as a high-performance anode material for rechargeable MMBs.

Conclusion

In this study, we demonstrated a rational design strategy for achieving a highly reversible Mg metal anode by embedding Zn single atoms into a porous carbon nanofiber scaffold (ZnSA@PCF). The nanoscale hollow carbon cages derived from metal–organic framework precursors further enhance the specific surface area of the 3D host, offering abundant Mg²⁺ active sites and lowering the local nucleation current density. The atomic dispersion of Zn not only prevents agglomeration but also provides ultra-uniform magnesiophilic sites that effectively facilitate homogeneous Mg nucleation. Zn single atoms do not induce structural distortion in the host and maintain their atomic dispersion without aggregation over extended cycling, enabling effective guidance of uniform Mg nucleation. Electrochemical performance evaluations confirmed the superior charge transfer kinetics and nucleation behavior of ZnSA@PCF. These findings not only validate the effectiveness of single-atom magnesiophilic agents in regulating interfacial Mg behavior but also demonstrate the strong potential of ZnSA@PCF as a practical anode platform for next-generation MMBs.

Author contributions

Jun-Won Lee and Jeong Ho Na: methodology, data analysis, validation, investigation, visualization, writing – original draft. SeongJae Lee: formal analysis, software, simulation. Seonju Kim: resources, formal analysis. Hee Seung Ryu: resources, visualization. Kyeounghak Kim: validation. Haeseong Jang: data curation, formal analysis. Seung-Keun Park: supervision, resources. Hee-Dae Lim: supervision, conceptualization, writing – review & editing.

Conflicts of interest

The authors declare no competing financial interest.

Data availability

All data supporting the findings of this study are available within the article and its SI files.

Supplementary information: detailed experimental procedures, additional electrochemical data, structural characterization results, and supporting discussions. See DOI: https://doi.org/10.1039/d5ta06095h.

Acknowledgements

This work was supported by the National Research Foundation of Korea grant funded by the Korean Government (MSIT) (RS-2024-00335171) and by the Industrial Technology Innovation Program (P268600148) of the Ministry of Trade, Industry and Energy of the Republic of Korea. This work was also supported by the Technology Innovation Program (2410001120) through the Korea Planning & Evaluation Institute of Industrial Technology (KEIT) funded by the Ministry of Trade, Industry & Energy (MOTIE, Korea). This research was supported by Plastic Free Specialized Graduate Program through the Korea Environmental Industry & Technology Institute (KEITI) funded by the Ministry of Environment (MOE). This research was supported by the Regional Innovation System & Education (RISE) through the Seoul RISE Center funded by the Ministry of Education (MOE) and the Seoul Metropolitan Government (2025-RISE-01-027-04). This work was financially supported by the Materials/Parts Technology Development Program (no. RS-2024-00456324) funded by the Ministry of Trade, Industry & Energy (MOTIE, Korea).

References

1. J. Kim, M. Kim, J. Lee, J. An, S. Yang, H. C. Ahn, D.-J. Yoo and J. W. Choi, Chem. Soc. Rev., 2024, 53, 8878–8902.
2. J. Niu, Z. Zhang and D. Aurbach, Adv. Energy Mater., 2020, 10, 2000697.
3. W. M. Haynes, CRC Handbook of Chemistry and Physics, CRC press, 2016.
4. R. Mohtadi, O. Tutusaus, T. S. Arthur, Z. Zhao-Karger and M. Fichtner, Joule, 2021, 5, 581–617.
5. H. Wang, J. Ryu, Y. Shao, V. Murugesan, K. Persson, K. Zavadil, K. T. Mueller and J. Liu, ChemElectroChem, 2021, 8, 3013–3029.
6. Y. Liang, H. Dong, D. Aurbach and Y. Yao, Nat. Energy, 2020, 5, 646–656.
7. Z. Liu, Y. Li, Y. Ji, Q. Zhang, X. Xiao, Y. Yao, L.-Q. Chen and Y. Qi, Cell Rep. Phys. Sci., 2021, 2(1), 100294.
8. M. Matsui, J. Power Sources, 2011, 196, 7048–7055.
9. M. Jäckle and A. Groß, J. Chem. Phys., 2014, 141.
10. M. Jäckle, K. Helmbrecht, M. Smits, D. Stottmeister and A. Groß, Energy Environ. Sci., 2018, 11, 3400–3407.
11. C. Ling, D. Banerjee and M. Matsui, Electrochim. Acta, 2012, 76, 270–274.
12. J. H. Kwak, Y. Jeoun, S. H. Oh, S. Yu, J.-H. Lim, Y.-E. Sung, S.-H. Yu and H.-D. Lim, ACS Energy Lett., 2021, 7, 162–170.
13. R. Davidson, A. Verma, D. Santos, F. Hao, C. Fincher, S. Xiang, J. Van Buskirk, K. Xie, M. Pharr and P. P. Mukherjee, ACS Energy Lett., 2018, 4, 375–376.
14. X. Liu, G. Wang, Z. Lv, A. Du, S. Dong and G. Cui, Adv. Mater., 2024, 36, 2306395.
15. M. S. Ding, T. Diemant, R. J. Behm, S. Passerini and G. A. Giffin, J. Electrochem. Soc., 2018, 165, A1983.
16. C. Li, R. D. Guha, A. Shyamsunder, K. A. Persson and L. F. Nazar, Energy Environ. Sci., 2024, 17, 190–201.
17. J. Wang, W. Zhao, H. Dou, B. Wan, Y. Zhang, W. Li, X. Zhao and X. Yang, ACS Appl. Mater. Interfaces, 2020, 12, 19601–19606.
18. Z. Zhao-Karger, R. Liu, W. Dai, Z. Li, T. Diemant, B. Vinayan, C. Bonatto Minella, X. Yu, A. Manthiram and R. J. r. Behm, ACS Energy Lett., 2018, 3, 2005–2013.
19. M. Zhang, W. Zhao, Y. Liu, M. Zhou, Z. Pan and X. Yang, ACS Energy Lett., 2025, 10, 552–561.
20. Y. Zhao, X. Zhang, J. Xiao, H. Fan, J. Zhang, H. Liu, Y. Liu, H. Yuan, S. Fan and Y. Zhang, ACS Appl. Mater. Interfaces, 2022, 14, 6499–6506.
21. G. Yang, Y. Li, J. Wang, Y. Lum, C. Y. J. Lim, M.-F. Ng, C. Zhang, Z. Chang, Z. Zhang and A. D. Handoko, Energy Environ. Sci., 2024, 17, 1141–1152.
22. D. Chinnadurai, W. Y. Lieu, S. Kumar, G. Yang, Y. Li and Z. W. Seh, Nano Lett., 2023, 23, 1564–1572.
23. J. Bae, H. Park, X. Guo, X. Zhang, J. H. Warner and G. Yu, Energy Environ. Sci., 2021, 14, 4391–4399.
24. C. Wei, L. Tan, Y. Zhang, B. Xi, S. Xiong, J. Feng and Y. Qian, Energy Storage Mater., 2022, 48, 447–457.
25. H. Park, H.-K. Lim, S. H. Oh, J. Park, H.-D. Lim and K. Kang, ACS Energy Lett., 2020, 5, 3733–3740.
26. D. Zhang, Y. Sun, X. Liu, Y. Zhang, R. Wang, Y. Zhao, M. Pan, Y. Wang, S. Chen and M. Zhou, ACS Energy Lett., 2024, 9, 2685–2695.
27. F. Wang, D. Wu, Y. Zhuang, J. Li, X. Nie, J. Zeng and J. Zhao, ACS Appl. Mater. Interfaces, 2022, 14, 31148–31159.
28. B. Wan, H. Dou, X. Zhao, J. Wang, W. Zhao, M. Guo, Y. Zhang, J. Li, Z.-F. Ma and X. Yang, ACS Appl. Mater. Interfaces, 2020, 12, 28298–28305.
29. Z. Song, Z. Zhang, A. Du, S. Dong, G. Li and G. Cui, Adv. Mater., 2021, 33, 2100224.
30. J. Liu, J. Zhang, Z. Zhang, A. Du, S. Dong, Z. Zhou, X. Guo, Q. Wang, Z. Li and G. Li, ACS Nano, 2022, 16, 9894–9907.
31. S. Lee, D. W. Kang, J. H. Kwak, S. Shin, J.-W. Park, S.-H. Yu, H.-G. Jung, B. G. Kim and H.-D. Lim, Chem. Eng. J., 2022, 431, 133968.
32. J. H. Kwak, S. Shin, Y. Jeoun, Y. Lee, S. Yu, Y. S. Yun, Y.-E. Sung, S.-H. Yu and H.-D. Lim, J. Power Sources, 2022, 541, 231724.
33. Y. Li, G. Yang, S. Sun, C. Zhang, C. Y. J. Lim, A. J. Y. Wong, W. Y. Lieu, Z. Sofer, M.-F. Ng and W. Liu, Nano Lett., 2022, 22, 6808–6815.
34. H.-D. Lim, D. H. Kim, S. Park, M. E. Lee, H.-J. Jin, S. Yu, S. H. Oh and Y. S. Yun, ACS Appl. Mater. Interfaces, 2019, 11, 38754–38761.
35. J. Bi, Y. Liu, Z. Du, K. Wang, W. Guan, H. Wu, W. Ai and W. Huang, Adv. Mater., 2024, 36, 2309339.
36. F. Li, H. Dou, Z. Zhao, W. Li, Q. Bai, S. Yang and X. Wang, Inorg. Chem. Front., 2025, 12, 1517–1527.
37. Q. Li, S. Zhu and Y. Lu, Adv. Funct. Mater., 2017, 27, 1606422.
38. S. Jin, Y. Jiang, H. Ji and Y. Yu, Adv. Mater., 2018, 30, 1802014.
39. J. H. Kwak, S. Park, S. Shin, S. Park, C. Kang, S.-H. Yu, J. Moon and H.-D. Lim, Energy Storage Mater., 2023, 59, 102762.
40. P. Zhai, T. Wang, W. Yang, S. Cui, P. Zhang, A. Nie, Q. Zhang and Y. Gong, Adv. Energy Mater., 2019, 9, 1804019.
41. Y. Sun, J. Zhou, H. Ji, J. Liu, T. Qian and C. Yan, ACS Appl. Mater. Interfaces, 2019, 11, 32008–32014.
42. S. Huang, S. Lu, Y. Lv, N. Li, Z. Wu, G. Zhong, X. Ren, Y. Wang, B. Sun and Y. Huang, Nano Res., 2023, 16, 11473–11485.
43. J. Wang, J. Zhang, S. Duan, L. Jia, Q. Xiao, H. Liu, H. Hu, S. Cheng, Z. Zhang and L. Li, Nano Lett., 2022, 22, 8008–8017.
44. G. Li, Y. Tang, Y. Wang, S. Cui, H. Chen, Y. Hu, H. Pang and L. Han, Angew. Chem., 2025, 137, e202417602.
45. X. Luan, L. Qi, Z. Zheng, Y. Gao, Y. Xue and Y. Li, Angew. Chem., Int. Ed., 2023, 62, e202215968.
46. Z. Lu, J. Wang, W. Feng, X. Yin, X. Feng, S. Zhao, C. Li, R. Wang, Q. A. Huang and Y. Zhao, Adv. Mater., 2023, 35, 2211461.
47. Q. Lai, Y. Zhao, Y. Liang, J. He and J. Chen, Adv. Funct. Mater., 2016, 26, 8334–8344.
48. R. Ahmad, U. A. Khan, N. Iqbal and T. Noor, RSC Adv., 2020, 10, 43733–43750.
49. S. Gadipelli and Z. X. Guo, ChemSusChem, 2015, 8, 2123–2132.
50. J. Chastain and R. C. King Jr, Handbook of X-Ray Photoelectron Spectroscopy, PerkinElmer Corporation40 1992.
51. Z. Hao, J. Chen, D. Zhang, L. Zheng, Y. Li, Z. Yin, G. He, L. Jiao, Z. Wen and X.-J. Lv, Sci. Bull., 2021, 66, 1649–1658.
52. G. Wei, Y. Li, X. Liu, J. Huang, M. Liu, D. Luan, S. Gao and X. W. Lou, Angew. Chem., 2023, 135, e202313914.
53. J. Zhao, X. Ren, X. Liu, X. Kuang, H. Wang, C. Zhang, Q. Wei and D. Wu, Chem. Eng. J., 2023, 452, 139533.
54. L. Han, S. Song, M. Liu, S. Yao, Z. Liang, H. Cheng, Z. Ren, W. Liu, R. Lin and G. Qi, J. Am. Chem. Soc., 2020, 142, 12563–12567.
55. Z. Liu, J. Ren, F. Wang, X. Liu, Q. Zhang, J. Liu, P. Kaghazchi, D. Ma, Z. Chi and L. Wang, ACS Appl. Mater. Interfaces, 2021, 13, 27085–27095.
56. J. Venables, Phys. Rev. B:Condens. Matter Mater. Phys., 1987, 36, 4153.
57. C. Hou, S. Ming, K. Chen, J. Rong, X. Yu, L. Yu and H. Hou, J. Phys. Chem., 2024, 128, 3204–3213.

---

Footnote

† Equally contributing authors.

---