Long-distance lanthanide migration regulated by interfacial lattice strain in nanostructures Authors: Yachong Liu, Xi Zou, Xinle Tian, Ruizhe Xiao, Zhiqiang Hu, Hao Jiang, Yan Su, Cheng Han, Xiyan Li, Xian Qin, Sanyang Han, Hongjie Zhang, Qianqian Su Published: 2026 ABSTRACT AbstractInterfacial strain engineering by controlling lattice mismatch holds implications for both fundamental studies and technological applications of heterostructured nanomaterials. However, defining its role remains challenging because the type, magnitude, and spatial distribution of strain are difficult to regulate at the nanoscale. Here, we establish a controllable model system to regulate interfacial lattice strain in lanthanide-doped core-shell-shell nanolattices by tuning lanthanide composition and shell deposition rate. We show that lattice mismatch governs lanthanide migration across heterointerfaces, thereby dictating crystal growth and optical properties. In hexagonal-phase NaErF4@NaYF4@NaLnF4 nanolattices, when the lattice mismatch exceeds 5.1%, lanthanide ions migrate by up to 13 nm from the core into the shell along the < 1000> crystallographic direction. Mechanistic studies reveal that large interfacial strain not only promotes anisotropic shell growth by triggering preferential deposition of shell precursors onto specific crystal surfaces but also facilitates the formation of lanthanide vacancies at the core-shell interface. These findings identify interfacial strain as a key regulator of ion diffusion and structural evolution in heterostructured nanocrystals, providing a design principle for constructing heterostructures with spatially confined active ions for applications in bioimaging, nanocatalysis, and quantum information.Subject terms: Nanoparticles, Nanoparticles Introduction Strain engineering has emerged as a crucial strategy for directing heteroepitaxial growth and modulating the optical, electrical, catalytic, and magnetic functionalities of advanced nanomaterials1–7. In heterogeneous core-shell systems, including metallic nanocrystals8, semiconductor quantum dots9, rare-earth nanocrystals10,11, and multicomponent complex nanostructures12, lattice mismatch between adjacent domains generates interfacial strain. Such strain strongly influences shell nucleation, growth orientation, interface coherence, and ultimately the morphology and properties of the resulting nanocrystals. However, interfacial strain is often associated with a high-energy, thermodynamically unstable state and can be partially alleviated through ion diffusion, defect formation, or phase transition, driving the system from a metastable state to a more stable configuration12–14. Even after partial relaxation, residual strain may continue to substantially impact the structural integrity of the epitaxial shell, interface stability, and local crystal symmetry, which in turn affects the overall properties of the resulting nanocrystals. Therefore, rational regulation of interfacial strain provides a powerful route to control lattice symmetry, chemical bonding, and electronic structures, enabling nanomaterials with tailored morphologies, interfaces, and functionalities. Lanthanide elements, featuring unique 4f electronic configurations and ladder-like energy levels, have become essential building blocks of next-generation optical functional materials, showing broad application prospects in defense, biomedicine, and information storage15–18. In lanthanide-doped nanocrystals, core-shell engineering is widely used to suppress surface quenching, spatially separate active ions, regulate energy transfer pathways, and enhance luminescence performance. Recently, lattice mismatch-induced interfacial strain has attracted increasing attention as an additional parameter for regulating lanthanide nanophosphors. Previous studies have shown that interfacial strain can strongly affect the morphological evolution of lanthanide nanocrystals19–21. Meanwhile, cation interdiffusion across core-shell interfaces has also been observed22–24. Furthermore, interfacial strain-mediated modulation of optical properties has been explored for advanced applications such as nanothermometry, deep-tissue bioimaging, and high-pressure probing21,25,26. These advances have laid the groundwork for breakthroughs in highly sensitive detection, high-performance nanophotonic devices, and bioimaging probes. Despite these advances, the mechanistic role of interfacial strain in lanthanide-doped core-shell nanocrystals remains insufficiently understood9,27. In particular, the correlation of lattice mismatch with interface integrity, crystal growth, defect formation, and ion migration has not been clearly established. Therefore, understanding the mechanistic effect of interfacial strain is essential for unlocking the full potential of core-shell nanoparticles, enabling the development of nanomaterials with tailored properties and advancing diverse applications in nanotechnology. Herein, we engineer interfacial strain in NaLnF4 (Ln = Lu, Ho, Y, Dy, Tb, Gd, Eu, Nd, Pr, Ce) core-shell-shell nanocrystals by selecting shell precursors with varying ionic radii and by modulating deposition kinetics. Using this platform, we demonstrate that interfacial lattice misfit is the primary driving force for long-distance lanthanide migration and the resultant morphology evolution. Mechanistically, lattice mismatch-induced strain both biases shell precursor deposition on specific crystallographic facets and lowers the formation energy of lattice defects at interfaces, collectively promoting the diffusion dynamics of lanthanide ions (Fig. 1). Schematic illustration of interfacial strain-driven lanthanide migration in anisotropically grown core-shell-shell nanocrystals. a Schematic representation of the directional epitaxial growth of the NaNdF4 outer shell on the β-NaYF4 intermediate shell. Lattice mismatch between the NaYF4 intermediate shell and the NaNdF4 outer shell differs at the (0001) and (1000) interfacial planes, generating anisotropic interfacial strain during shell growth. Here, a and c denote the lattice parameters of the β-NaLnF4 crystals, aY and cY denote those of the NaYF4 intermediate shell, aNd and cNd denote those of the NaNdF4 outer shell, and ξa and ξc denote the lattice mismatches along the a and c directions, respectively. b Schematic illustration of strain-driven lanthanide migration from the core towards the outer shell. Large interfacial lattice misfit at the metastable heterointerface promotes long-distance cation migration and is accompanied by morphology evolution during anisotropic shell growth. Results Design and synthesis of NaErF4@NaYF4@NaNdF4 As a proof of concept, we established a NaErF4@NaYF4@NaNdF4 core-shell-shell nanostructure as a model system to elucidate the interplay between interfacial strain, anisotropic growth, and lanthanide migration. This design relies on a synergistic combination of structural and optical features. Structurally, the system is engineered to isolate strain effect. The negligible lattice mismatch at the inner NaErF4-NaYF4 interface ensures coherent initial growth, while the significant mismatch at the outer NaYF4-NaNdF4 interface introduces the strain necessary to drive facet-dependent growth and cation diffusion (Supplementary Tables 1-4). Optically, the inert NaYF4 spacer prevents direct intermixing between Er3+ and Nd3+. As a result, the well-separated, spectrally-distinct Er3+ and Nd3+ emitters enable the use of photoluminescence dynamics as a sensitive probe for strain-mediated lanthanide migration. We studied the influence of interfacial strain on both the morphological evolution of the multilayered nanocrystals and the migration of lanthanides by monitoring their spatial distribution (Fig. 2a). To establish a baseline, we synthesized NaErF4@NaYF4 core-shell nanoparticles (Supplementary Table 3). Transmission electron microscopy (TEM) images showed that a NaYF4 shell exhibited a seamless coating on the NaErF4 core, which is ascribed to a negligible mismatch (~0.5%) of core and shell lattice (Supplementary Fig. 1). Energy dispersive X-ray (EDX) elemental mapping analysis indicated an even distribution of Er3+ and Y3+ ions in the core and shell regions, respectively (Fig. 2b). Long-range lanthanide migration. a Schematics showing the growth of NaErF4@NaYF4@NaNdF4 (Er@Y@Nd) nanostructures. b,c TEM images, high-resolution TEM images (inset), and EDX elemental mapping of NaErF4@NaYF4 (Er@Y) and NaErF4@NaYF4@NaNdF4 nanocrystals (Injection rate: 6.7 μmol/min). d HAADF-STEM image and corresponding false-color maps of relative d-spacing in the {101¯1} planes of NaErF4@NaYF4@NaNdF4 nanocrystals (Injection rate: 6.7 μmol/min). e,f Emission spectra and luminescence decay curves of NaErF4@NaYF4 and NaErF4@NaYF4@NaNdF4 nanocrystals. Injection rate: 6.7 μmol/min. g TEM images, HAADF-STEM image, and EDX elemental mapping of NaErF4@NaYF4@NaYF4 (Er@Y@Y) nanocrystals. Injection rate: 6.7 μmol/min. We next coated a NaNdF4 shell on the preformed NaErF4@NaYF4 core-shell nanoparticle by hot injection at a low injection rate of approximately 6.7 μmol/min (injection time: 60 min). TEM images revealed that the resulting multilayered NaErF4@NaYF4@NaNdF4 nanoparticles crystallized in a rod-like morphology featuring a pronounced central necking (Fig. 2c). We found that Y3+ ions were not evenly wrapped around the NaErF4 cores, but diffused to the two ends of the nanorods, and distributed uniformly in the NaNdF4 shell. Interestingly, elemental mapping showed a substantial amount of Er3+ ions diffusing across the interfaces to the NaNdF4 layer (Fig. 2c), along with the intermixing between Y3+ and Nd3+ ions in the shell layers. The results also show that both Y and Er signals are highest at their original locations and gradually decrease toward the two ends of the nanorods (Supplementary Fig. 2). Surprisingly, a long-distance migration of approximately 13 nm along the <0001> direction was estimated in the hetero-nanostructures under study. To map the spatial distribution of lattice misfit strain at the atomic scale, we performed geometric phase analysis (GPA) on High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) images, focusing on the {101¯1} planes. The resulting strain maps reveal that the NaNdF4 shell possesses a larger lattice spacing than the NaErF4@NaYF4 seed (Fig. 2d). This lattice mismatch results in a steep strain gradient at the NaYF4-NaNdF4 interface. Specifically, the NaNdF4 lattice exerts tensile strain on the underlying NaYF4 interlayer, while the shell itself experiences compressive stress (Supplementary Fig. 3). Collectively, these results provide direct, atomic-level evidence for the strain distribution within the NaErF4@NaYF4@NaNdF4 heterostructure. The lanthanide intermixing significantly shortened the distance between the Nd3+ and Er3+ ions, which could trigger strong resonant energy transfer from the Er3+ to Nd3+ ions upon light excitation28,29. By directly pumping the Er3+ emitters, the spectroscopy measurements showed markedly quenched photoluminescence and shortened lifetime of Er3+ in the NaErF4@NaYF4@NaNdF4 nanoparticles compared with that observed in the NaErF4@NaYF4 nanoparticles, again confirming the close proximity of Nd3+ to Er3+ ions and the reduced structural integrity of the shell caused by large interfacial lattice misfit (Fig. 2e, f). Morphological evolution investigation We ruled out dissolution-recrystallization of NaErF4@NaYF4 as the primary driver for the observed central necking and cation diffusion during the NaNdF4-shell growth through lattice-matched controls and kinetics studies30. First, to test chemical stability, we coated NaErF4@NaYF4 seeds with a lattice-matched NaYF4 shell under identical conditions (6.7 μmol/min). In contrast to the strained NaNdF4 system, this isostructural growth yielded a uniform shell with no reduction in the seed’s axis, while the NaErF4 core morphology remains unaltered and Er3+ ions are strictly confined to the core region (Fig. 2g and Supplementary Fig. 4). These findings confirm that the synthetic environment is not inherently corrosive and favors deposition over dissolution. Second, we further investigated the kinetic dependence of the transformation by annealing NaErF4@NaYF4@NaNdF4 nanoparticles prepared at different injection rates (80 vs. 6.7 μmol/min) at 310 °C for 60 min. The sample synthesized rapidly (80 μmol/min) underwent a dramatic star-to-rod transition with partial Er3+ migration, whereas those synthesized slowly exhibited only mild structural changes (Supplementary Figs. 5 and 6). This pronounced rate dependence indicates that the main driving force is the relaxation of accumulated interfacial strain, which scales with growth rate, rather than thermodynamic solubility. Furthermore, the absence of separate, smaller crystallites typically associated with classical dissolution-recrystallization reinforces a mechanism of strain-induced mass transport to lower the system’s Gibbs free energy30. To quantify the contribution of lanthanide dissolution to the growth mechanism, we performed time-resolved monitoring of the NaNdF4 shell deposition on the surface of NaErF4@NaYF4 seeds (~6.7 μmol/min; 60 min total). TEM analysis of samples taken at 1, 5, 10, 20, 30, and 60 min reveals a continuous morphological evolution from quasi-spherical to quasi-rhombus, and ultimately to rod-like particles with a distinct central necking (Supplementary Figs. 7 and 8). This progressive transformation, devoid of secondary nucleation, suggests anisotropic growth rather than Ostwald ripening. We corroborated this by determining the concentration of free lanthanide ions in the reaction supernatant via time-resolved inductively coupled plasma-optical emission spectrometry (ICP-OES). Throughout the reaction, Y3+ and Er3+ levels remained negligible (≤0.01% and ≤0.003% of their total content, respectively) (Supplementary Table 5). Additionally, we examined the possibility of ligand-assisted dissolution by annealing NaErF4@NaYF4 seeds (0.4 mmol) in the presence of oleic acid and ODE at 310 °C for 60 min. Even under such harsh conditions, only 0.43% of Er3+ and 1.91% of Y3+ were detected (Supplementary Table 6). Although the slightly elevated Y3+ release suggests localized destabilization of the intermediate layer, these quantities are insufficient to account for the massive ion redistribution observed in NaErF4@NaYF4@NaNdF4 nanorods. Collectively, these data confirm that the morphological reshaping and lanthanide migration arise predominantly from strain-mediated solid-state diffusion rather than dissolution-recrystallization process. To elucidate the mechanism underlying the anisotropic growth from an atomic perspective, we calculated the surface energies of the two primary low-index facets using density functional theory (DFT). The results revealed that the (0001) facet possesses a lower surface energy than the (1000) facet (Supplementary Fig. 9 and Supplementary Table 7). Based on classical thermodynamic considerations, one would therefore expect the ripening process to favor deposition onto the (1000) facet to minimize its surface energy21. However, our experimental observation contradicts this expectation, indicating that this paradox might be governed by the substantial lattice mismatch between NaYF4 and NaNdF4 lattices. To verify it, we simulated the adsorption of an Nd atom onto the surfaces of pure NaYF4, pure NaNdF4, and heterostructured NaYF4-NaNdF4 to probe the impact of interfacial strain (Supplementary Figs. 10 and 11). The calculated adsorption energies demonstrated a shift: while Nd adsorption is favored on the (1000) surface of pure NaNdF4, the presence of interfacial strain redirects this preference towards the (0001) surface (Supplementary Figs. 12, 13, and Supplementary Table 8). This reversal indicates that, in this system, interfacial strain exerts a more dominant influence than inherent surface energy in regulating the direction of epitaxial growth. Collectively, the combined experimental and computational results provide energetic and kinetic support for a strain-driven preferential deposition and intraparticle diffusion mechanism. Note that the aforementioned shell growth was achieved at a relatively low injection rate of ~6.7 μmol/min, which suggests that the pre-deposited shell precursors have sufficient time to reach a thermodynamically more stable configuration through intraparticle diffusion processes before the next wave of precursor deposition31. In this regard, it is highly possible to modulate lanthanide diffusion through the control over meta-states via the adjustments to precursor injection rate (Supplementary Fig. 14). For the NaNdF4 shells grown at different injection rates, X-ray diffraction (XRD) patterns confirmed that all samples maintained a pure hexagonal crystal structure (Supplementary Fig. 15). The diffraction peaks of the multilayered structures shifted to lower diffraction angles relative to the NaErF4@NaYF4 nanocrystals pattern, demonstrating an incoherent epitaxial growth with expanded seed lattice19. In addition, the asymmetrical broadening of these diffraction peaks indicates an uneven distribution of lattice spacings, likely due to the formation of lattice defects or distorted interfaces in the heterostructures. TEM images revealed dumbbell-like NaNdF4 shell fragments on the NaErF4@NaYF4 nanocrystals at an injection rate of 80.0 μmol/min over 5 min (Fig. 3a). As the injection rate decreased, the nanoparticle morphology evolved from a star-like (40.0 and 20.0 μmol/min) to a triangular (13.3 μmol/min) and finally to a rod-like shape (6.7 μmol/min) (Figs. 3b–d, 2c). Moreover, the short axis of the NaErF4@NaYF4@NaNdF4 nanorods progressively decreased with decreasing precursor injection rate (Supplementary Fig. 16). These results indicate that variations in the injection rate affect the growth dynamics of NaNdF₄ shells. This behavior aligns with previous studies, which found that a fast-sequential injection of the shell precursor resulted in isotropic shell growth, while a slow dropwise injection led to rod-like shell morphologies in NaYF4@NaYF4 core-shell nanoparticles20. The crystal structure and lattice mismatch in nanocrystals are visualized through high-resolution TEM (HRTEM) characterization. HRTEM also showed that NaErF4@NaYF4 nanocrystals initially exhibited a single crystal structure (Supplementary Fig. 17). Upon introducing Nd3+ ions, significant structural changes were observed (Supplementary Figs. 18–20). At high injection rates, the nanocrystals displayed a polycrystalline structure (Fig. 3e, f) and transformed to a single crystal-like structure at slow injection rates (Fig. 3g, h). We also examined the interfacial regions using HRTEM and inverse fast Fourier transform (IFFT) analysis and found discontinuities in lattice fringes (Fig. 3e, f), indicative of lattice-mismatch-induced structural distortion, while lattice fringes became more orderly at lower injection rates (Fig. 3g, h and Supplementary Fig. 21). Shell growth dynamics and strain relaxation modulated by the Nd-shell precursor injection rate. a–d TEM images of NaErF4@NaYF4@NaNdF4 nanocrystals synthesized with Nd-shell injection rates of 80.0, 40.0, 20.0 and 13.3 μmol/min, respectively. As the injection rate decreases, shell growth progressively approaches a thermodynamically more stable state. e–h HRTEM images and corresponding inverse fast Fourier transform (IFFT) analyses of the nanoparticles. Regions E, F, G and H are enlarged views of the distorted lattice fringes near the heterointerface. In inset E, the dotted lines mark the interface boundaries. E1–H1 show the corresponding IFFT patterns, in which T indicates the trend of interfacial strain. i–l EDX elemental mapping and corresponding atomic models showing lanthanide migration in the nanocrystals prepared at the same injection rates as in (a–d). Together, these results show that decreasing the precursor injection rate promotes long-range lanthanide migration and strain relaxation, driving morphology evolution towards a more rod-like core-shell-shell structure. With the rapid injection of the Nd precursor, the Er core retains its spherical shape while the distribution of Y3+ ions originally located in the intermediate layer begins to resemble that of the outer Nd shell (Fig. 3i, j and Supplementary Fig. 22a, b). In contrast, at low injection rates, the shape of Er core increasingly mimics that of the Nd shell, and the distribution of Y3+ ions tends to overlap with the Nd shell (Fig. 3k, l and Supplementary Fig. 22c, d). These results suggest that high injection rates lead to uneven precursor distribution and confined growth, limiting lanthanide diffusion to thermodynamically stable sites to some extent. This arises because the pre-deposited shell precursors do not have sufficient time to release the interfacial strain via ion diffusion. In contrast, low injection rates facilitate uniform distribution of precursors, allowing pre-deposited ions to diffuse toward stable sites sufficiently before the arrival of the next batch of precursors. Thus, the interplay between injection rate, growth kinetics, and surface energy critically determines the morphological changes of the nanocrystals and diffusion of lanthanides. We found that the luminescence from Er3+ emitters was markedly attenuated as the injection rates decreased (Supplementary Fig. 23). For instance, nanoparticles synthesized at an injection rate of 40.0 μmol/min exhibited approximately 29-fold and 100-fold higher luminescence at 500–700 nm and 1400–1600 nm, respectively, compared with that recorded for the nanoparticle prepared at an injection rate of 6.7 μmol/min (Supplementary Fig. 24). Note that the corresponding luminescence lifetime gradually decreased as the injection rate decreased. These findings corroborated enhanced energy transfer from the Er3+ ions to their neighboring Nd3+ ions and strong lanthanide intermixing in these nanoparticles. To evaluate the influence of interfacial strain on lanthanide diffusion length, we synthesized a class of NaErF4@NaYF4@NaNdF4 nanocrystals with varying NaYF4 spacer thicknesses. Er3+ ions migrated from the core to the outer NaNdF4 shell even when the thickness of the Y-spacer was up to 6.8 nm (Supplementary Fig. 25). This observation demonstrates that lattice misfit strain at the outer Nd-Y interface drives a long-range relaxation dynamic capable of mobilizing ions over substantial physical distances. Of note, even a thin outer NaNdF4 layer proved sufficient to generate the interfacial strain required to trigger this migration (Supplementary Fig. 26). To further quantify the relationship between strain magnitude and ion transport, we reduced the lattice mismatch by synthesizing nanoparticles with a mixed-composition outer shell (NaErF4@NaYF4@NaNdF4:50%Y). The resulting nanoparticles presented a more regular nanorod morphology. Specifically, the NaYF4 interlayer remained largely intact, and Er3+ ions are predominantly confined to the core with only limited outward migration (Supplementary Fig. 27). While the reduced Nd3+ concentration still generates enough strain to induce minor migration, it is insufficient to drive the severe Y3+ depletion or partial shell removal observed with a pure NaNdF4 shell. These comparative results confirm that the extent of lanthanide migration and morphological evolution is directly governed by the magnitude of the interfacial lattice mismatch. Mechanistic investigation of lanthanide migration The reason why interfacial lattice mismatch can induce lanthanide migration and redistribution in multilayered nanocrystals is that lattice vacancies create empty sites within the crystal structure, which can serve as pathways for ions to move. In this regard, the rate of substitutional diffusion is influenced by two primary factors: the ease of vacancy formation and the mobility of atoms into those vacancies32,33. Interfacial misfit strain in the NaErF4@NaYF4@NaNdF4 core-shell system is expected to promote lanthanide diffusion as the substantial lattice mismatch between NaYF4 and NaNdF4 introduces structural distortions that facilitate vacancy creation (Fig. 4a). The influence of interfacial strain on lanthanide vacancy formation and lanthanide migration. a A schematic diagram showing vacancy-mediated lanthanide diffusion. b Atomic illustration of lanthanide vacancies in a Ln-Ln-Ln chain of β-NaLnF4. Lni (i = 1–3) denotes Ln ions site in the surface, near-surface, and interface regions, respectively. Fluorine ions are omitted for clarity. Note that the c axis is along the <0001> crystallographic direction. c Calculated formation energy of lanthanide vacancies (VLn) in NaYF4 and NaNdF4 systems under strained and unstrained conditions. d Calculated energy difference (ΔE) between an initial state with a vacancy at the interface and a final state where a neighboring Y/Nd atom has migrated to fill it. Y@3Nd and Y@1Nd represent three-layer and one-layer of NaNdF4 are deposited on the (0001) facet of the NaYF4 substrate. To elucidate the role of interfacial strain on defect formation, we calculated the formation energy of lanthanide vacancies (VLn) in NaYF4 and NaNdF4 systems under strained and unstrained conditions (Supplementary Fig. 28). Of note, we focused on lanthanide vacancies in a Ln-Ln-Ln chain, as prior work indicates that they are easier to form than in the adjacent Ln-Na-Ln chain (Fig. 4b)34. The calculated formation energies reveal that VLn formation is favorable in the surface and near-surface regions, as evidenced by the negative formation energies, irrespective of interfacial strain (Fig. 4c and Supplementary Table 9). While the interfacial strain exerts a negligible overall influence, a detailed analysis reveals critical, site-specific effects. Specifically, compressive strain lowers the formation energy of the surface VNd1 by 0.1 eV, facilitating its formation. Tensile strain enhances the formation of Y vacancies at the core-surface interface. For example, the formation energy of VY3 decreases by 0.626 eV upon expansion. These findings indicate that the interfacial strain selectively lowers the formation energy of lanthanide vacancies to some extent, which could facilitate lanthanide diffusion in heterostructured nanosystems. We next investigated the impact of interfacial strain on vacancy-mediated Ln diffusion by analyzing the calculated energy difference (ΔE) between an initial state with a vacancy at the interface and a final state where a neighboring Y atom has migrated to fill it (Fig. 4d and Supplementary Fig. 29). Although this ΔE is not the absolute diffusion barrier, it serves as a robust qualitative descriptor for comparing diffusion trends: a positive ΔE indicates an endothermic process, and its magnitude provides a lower-bond estimate for the forward energy barrier. For the case of strain-free NaYF4, the resulting ΔE values reveal that the diffusion of the near-surface Y into the surface VY1 (ΔE1) is more facile than the diffusion of the interior Y (governed by ΔE2), as manifested by the smaller ΔE1 (Fig. 4d and Supplementary Table 10). A similar trend holds for the strain-free NaNdF4 system. Critically, the introduction of strain alters these energy landscapes. When the NaNdF4 layer is compressively strained to the NaYF4 lattice, we found a slight decrease in ΔE1, indicating an easier surface-level diffusion. More significantly, when the NaYF4 is expanded, the ΔE2 for interior atom diffusion drops by 31.67%. This substantial reduction demonstrates that ensile strain markedly facilitates the migration of interior Y atoms towards VNd in the shell, which is consistent with the experimentally observed diffusion of Er3+ and Y3+ towards outer shells. Universality of strain-driven lanthanide migration By taking advantage of lanthanide contraction, we studied the effects of interfacial strain and epitaxial growth dynamics on lanthanide diffusion. Specifically, the NaErF4@NaYF4 nanoparticles were coated with Lu, Ho, Y, Dy, Tb, Gd, Eu, Pr, and Ce shells, generating misfit strains of −2.0%, −0.06%, 0%, 0.7%, 1.4%, 2.0%, 2.4%, 6.0%, and 7.0%, respectively (Supplementary Tables 1–4). TEM images revealed that minor misfit strain on the NaErF4@NaYF4 surface promoted a coherent epitaxial growth, while higher misfit strain levels induced non-coherent growth (Fig. 5a and Supplementary Figs. 30–34), which is consistent with the previous report20,21. In the case of NaErF4@NaYF4@NaYF4 nanoparticles, the outermost Y-shell, which applied zero strain, maintained a single-crystal structure, with minimal Er3+ ion migration (Fig. 2g, and injection rate: 6.7 μmol/min). In the systems with negligible lattice mismatch, specifically NaHoF4 (−0.06%) and NaDyF4 (0.7%), the nanoparticles retained their morphology, and lanthanide ions remained strictly confined, without appreciable migration of Er3+ or redistribution of Y3+ (Fig. 5a, c, and Supplementary Figs. 35 and 36). In contrast, when applying compressive strain on the Er core through the coating of Lu shell, we observed that the strain barely affected the diffusion of Er3+ but did promote Y3+ diffusion, as evidenced by the similarity of its distribution to that of the outer Nd shell under slow growth conditions (Fig. 5b). Generality of strain-driven lanthanide migration in NaErF4@NaYF4@NaLnF4 core-shell-shell nanocrystals. a TEM images of NaErF4@NaYF4@NaLnF4 (Ln = Lu, Ho, Dy, Tb, Gd, Eu, Pr and Ce) nanocrystals synthesized with a NaLnF4 outer-shell injection rate of ~6.7 μmol/min. Under slow precursor injection, all particles evolve towards a rod-like morphology. b–g Representative EDX elemental mapping of NaErF4@NaYF4@NaLnF4 nanocrystals with Ln = Lu, Dy, Tb, Eu, Pr and Ce synthesized under the same conditions. When the outer shell contains Tb, Eu, Pr or Ce, the shell is under tensile strain, and the Er and Y signals show pronounced outward migration from the inner regions towards the outer shell. By contrast, when the outer shell contains Lu or Dy, Er and Y remain largely confined to the inner regions, indicating negligible outward migration. Interestingly, a Tb shell, with a modest misfit strain of 1.4%, triggered a significant long-range Er3+ migration under slow growth rates (Fig. 5d, injection rate: 6.7 μmol/min), suggesting that tensile strain plays a primary role in governing diffusion dynamics. In the case of Eu, Pr, and Ce shells, the large lattice mismatches at the Y-Eu, Y-Pr, and Y-Ce interfaces impeded uniform shell formation. These systems exhibited rod-like morphologies with pronounced central necking and substantial Er3+ diffusion at a low growth rate of 6.7 μmol/min (Fig. 5e–g). Despite the largest ionic radius of Ce3+, Er3+ diffusion could also be largely suppressed in the NaErF4@NaYF4@NaCeF4 when a higher growth rate was used (20.0 μmol/min), consistent with observations for Nd shells (Supplementary Fig. 37). Taken together, it is rational to infer that both the type and magnitude of interfacial strain play crucial roles in regulating morphological evolution and lanthanide diffusion, with large tensile strain effectively driving lanthanide diffusion over a long distance. Significant lanthanide diffusion was detected with Tb, Eu, Gd, Nd, Pr, and Ce coatings that impose tensile strain on the core, while Y, Ho, and Dy-coated nanoparticles exhibited almost no diffusion. Slow epitaxial growth facilitated lanthanide diffusion, whereas shell-imposed compressive strain and rapid shell growth could suppress lanthanide diffusion to some extent. The photoluminescence and lifetimes of the nanocrystals correlated with the distribution of the lanthanide ions (Supplementary Figs. 38–41). Discussion In summary, we investigated the effect of lattice mismatch-induced interfacial strain on heterogeneous epitaxial growth and ion diffusion within nanolattices. Using hexagonal-phase multilayered NaErF4@NaYF4@NaLnF4 (Ln = Lu, Ho, Y, Dy, Tb, Gd, Eu, Nd, Pr, and Ce) nanocrystals as model systems, we demonstrated that lanthanide diffusion across interfaces, which can be effectively controlled through misfit strain engineering, involving adjustments of shell growth dynamics and the type and magnitude of misfit strain (−2.0–7.0%). Specifically, in NaErF4@NaYF4@NaNdF4 nanocrystals with a substantial misfit strain of 5.1%, we observed that Er3+ ions in the core and Y3+ ions in the spacer migrated over 13 nm to the outer Nd shell, mainly along the <0001> crystallographic direction. These findings not only enhance our understanding of interfacial strain relaxation dynamics but also enable precise control over ion diffusion in heterostructure nanolattices, advancing the development of strain-engineered nanocrystals with intriguing electronic, optical, magnetic, or catalytic properties. Methods Materials Y(CH3CO2)3·xH2O (99.9%), Nd(CH3CO2)3·xH2O (99.9%), Gd(CH3CO2)3·xH2O (99.9%), Tb(CH3CO2)3·xH2O (99.9%), LuCl3·xH2O (99.9%), Er(CH3CO2)3·xH2O (99.9%), Ce(CH3CO2)3·xH2O (99.9%), Eu(CH3CO2)3·xH2O (99.9%), Pr(CH3CO2)3·xH2O (99.9%), Dy(CH3CO2)3·xH2O (99.9%), Ho(CH3CO2)3·xH2O (99.9%), NaOH (>98%), NH4F ( > 98%), oleic acid (OA, 90%), 1-octadecene (ODE, 90%), sodium oleate and sodium trifluoroacetate (Na-TFA) were all purchased from Sigma-Aldrich. Hydrochloric acid (AR), absolute methanol (AR), cyclohexane (AR) and ethanol (AR) were obtained from Sinopharm Chemical Reagent Co., Ltd. All chemicals were used as received without further purification. Synthesis of NaErF4 core nanocrystals NaErF4 core nanocrystals were prepared by a modified literature procedure35. A total of 4 mL of rare-earth acetate Er(CH3CO2)3 (0.275 g, 0.80 mmol) aqueous solution, oleic acid (10 mL), and octadecene (10 mL) were added to a 50 mL three-neck flask, and the mixture was stirred and heated to 150 °C for 1.5 h, and then cooled down to room temperature naturally. Next, the flask was transferred to the heating mantle, and then sodium oleate (0.761 g, 2.50 mmol) was added. The reaction was maintained at 100 °C under vacuum condition for 1 h. Subsequently, NH4F powder (0.148 g, 4.00 mmol) was added to the three-neck flask under nitrogen, and then slowly heated to 160 °C for 1.5 h. After vacuumizing for 10 min, the reaction was heated to 320 °C and kept at this temperature under nitrogen protection for 0.5 h before cooling down to room temperature. The resulting mixture was centrifugated and washed three times (20,000 g, 5 min) using ethanol/cyclohexane (v/v = 3/2), and finally dispersed in 8 mL cyclohexane prior to being used for shell coating. The synthesis of NaNdF4 and NaErF4:50%Nd was identical to that for NaErF4 nanoparticles except for using the different compositions of dopants. Synthesis of NaErF4@NaYF4 core-shell nanocrystals The core-shell nanocrystals were synthesized by a modified method11,35. The obtained NaErF4 core nanocrystals were used as seeds for the shell coating. A 2 mL aqueous solution of Y(CH3CO2)3 (0.106 g, 0.40 mmol), oleic acid (4 mL) and octadecene (6 mL) were added to a 50 mL three-neck flask. The temperature was heated to 150 °C and maintained at this temperature for 1.5 h before cooling down to 50 °C. Then, 4.4 mL methanol solution of NH4F (0.050 g, 1.36 mmol) and NaOH (0.040 g, 1.00 mmol) was quickly added and kept for 0.5 h, and then slowly heated to 80 °C for 0.5 h to remove methanol. At the same time, NaErF4 (4 mL) dispersed in cyclohexane was added into a 50 mL three-neck flask, then oleic acid (4 mL) and octadecene (6 mL) were added and heated to 80 °C to remove cyclohexane. Then, the temperature was slowly increased to 100 °C for 30 min in vacuo. After the end, the temperature was heated at a constant rate to 310 °C in a nitrogen environment. The treated shell precursor was injected at a uniform rate for 40.0 μmol/min (10 min), and then the reaction was cooled down to room temperature. The sample was centrifugally washed three times (20,000 g, 5 min) using ethanol/cyclohexane (v/v = 3/2), and finally dispersed in 4 mL cyclohexane. Synthesis of NaErF4@NaYF4@NaLnF4 core-multishell nanocrystals The preparation of the above-mentioned and other core-multishell nanocrystals, including NaErF4@NaYF4@NaNdF4 (Er@Y@Nd), NaErF4@NaYF4@NaTbF4 (Er@Y@Tb), NaErF4@NaYF4@NaYF4 (Er@Y@Y), NaErF4@NaYF4@NaGdF4 (Er@Y@Gd), NaErF4@NaYF4@NaLuF4 (Er@Y@Lu), NaErF4@NaYF4@NaCeF4 (Er@Y@Ce), NaErF4@NaYF4@NaPrF4 (Er@Y@Pr), NaErF4@NaYF4@NaEuF4 (Er@Y@Eu), NaErF4@NaYF4@NaDyF4 (Er@Y@Dy, 6.7 μmol/min), NaErF4@NaYF4@NaHoF4 (Er@Y@Ho, 6.7 μmol/min), were the same as that for core-shell nanocrystals except for the use of a shell with different lanthanide ions and shell injection rates. The preparation of NaErF4@NaYF4@NaNdF4 (6.7 μmol/min, sampling time: 1−60 min) core-multishell nanocrystals were the same as that for core-shell nanocrystals except for 0.5 mL of stock solution was taken at fixed time points. Finally, each sample was dispersed in 200 μL of cyclohexane for later use. The annealing treatment of NaErF4@NaYF4@NaNdF4 core-multishell nanocrystals NaErF4@NaYF4@NaNdF4 (4 mL) dispersed in cyclohexane was added into a 50 mL three-neck flask, then oleic acid (4 mL) and octadecene (6 mL) were added and heated to 80 °C to remove cyclohexane. Then, the temperature was slowly increased to 100 °C for 30 min in vacuo. After the end, the temperature was heated at a constant rate to 310 °C in a nitrogen environment, and kept at this temperature under nitrogen protection for 60 min before cooling down to room temperature. The sample was centrifugally washed three times (20,000 g, 5 min) using ethanol/cyclohexane (v/v = 3/2), and finally dispersed in 4 mL cyclohexane. Synthesis of NaErF4@NaYF4@NaNdF4 (thin NaNdF4 layer) core-multishell nanocrystals The synthetic procedures of NaErF4@NaYF4@NaNdF4 (thin NaNdF4 layer) nanocrystals were the same as that for core-shell nanocrystals except for the use of the 0.20 mmol Nd(CH3CO2)3 shell and shell injection rates of 6.7 μmol/min. Synthesis of NaErF4@NaYF4@NaYF4@NaNdF4 (80.0 μmol/min and 6.7 μmol/min, d(Y@Y) = 6.8 nm) core-multishell nanocrystals Firstly, NaErF4@NaYF4 core-shell nanocrystals were synthesized by the same method as above. NaErF4@NaYF4@NaYF4 (d(Y@Y) = 6.8 nm) core-shell nanocrystals were synthesized by a modified one-pot sequential layer-by-layer epitaxial growth method of the literature36. The whole synthesis process is mainly divided into two steps. The first is the precursor processing process (Refer to steps 1 and 2), and the second is the shell growth process (Refer to step 3). (Step 1) Treatment of Y-OA (0.1 M) shell precursor Rare-earth acetate powder Y(CH3CO2)3 (0.665 g, 2.5 mmol), ODE (15 mL), and OA (10 mL) were mixed and added into a 50 mL three-neck flask. The mixture solution was heated to 140 °C under vacuum and magnetic agitation for 30 min to remove water and oxygen. Finally, obtain a clear precursor solution Y-OA (0.1 M). (Step 2) Na-TFA-OA precursor solution Na-TFA (0.544 g, 4.0 mmol) and OA (10 mL) were added to a 50 mL three-neck flask under vacuum at room temperature while maintaining magnetic agitation to remove water and oxygen. Finally, colorless and transparent precursor solution Na-TFA-OA (0.4 M) was obtained. (Step 3) Synthesis of NaErF4@NaYF4@NaYF4 (d(Y@Y) = 6.8 nm) core-shell nanocrystals The cyclohexane solution of NaErF4@NaYF4 nanocrystals (2 mL), ODE (6 mL), and OA (4 mL) were added into a 50 mL three-neck flask. Under magnetic stirring, the mixture was raised to 80 °C, and kept for 20 min to remove cyclohexane. Then, the temperature was raised to 100 °C in vacuo for 20 min to remove methanol, oxygen and water from the reaction system. Next, Na-TFA-OA (0.4 M) and Y-OA (0.1 M) were injected alternately under nitrogen protection at a volume ratio of 1:2. The reaction was kept for 15 min before cooling down to room temperature. The products were centrifuged and washed three times (20,000 g, 5 min) with the mixture of ethanol and cyclohexane (v:v = 3:2). Finally, the sample was dispersed in 2 mL cyclohexane. Then, the NaErF4@NaYF4@NaYF4 nanocrystals were used as core nanoparticles, and the outermost layer was injected with Nd shell. The synthesis of the NaErF4@NaYF4@NaYF4@NaNdF4 (80.0 μmol/min and 6.7 μmol/min) nanocrystals was identical to that for core-shell nanocrystals except for the Nd shell injection rates. Physical measurements Transmission electron microscopy (TEM) measurements were carried out with an HT7700 field emission TEM operated at an acceleration voltage of 120 kV. Powder X-ray diffraction (XRD) results were collected on a Rigaku D/MAX-2200 attached with a rotating anode and a Cu Kα radiation source (λ = 0.15418 nm). The inductively coupled plasma (ICP) measurements were conducted by the SPECTRO ICP-OES (SPECTRO-BLUE). Luminescence spectra and decay curves measurements were performed with an FS5 (Edinburgh) combined with 980 nm (MDL-III-980) diode lasers at room temperature. High-resolution transmission electron microscopy (HRTEM) images, high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) images, Energy dispersive X-ray (EDX) elemental mapping and line scan analysis were carried out on JEM-ARM300F (300 kV) transmission electron microscope (Japan Electron Optics Laboratory/JEOL, ETA spherical aberration corrector). Unless otherwise noted, all spectra were measured under the same experimental conditions. In addition, all the experiments were repeated at least three times, and different experimental units (devices) were used. Supplementary information Supplementary Information Transparent Peer Review file Acknowledgements We thank the Key Laboratory of Biomedical Imaging Science and System, Chinese Academy of Sciences and Fujian Provincial Key Laboratory of Photonics Technology, Fujian Normal University for providing the facility and funding support. Author contributions Q.S. conceived the project. Q.S., X.Q., S.H., and H.Z. designed the experiments. X.Q. performed the computation calculations. Y.L., X.Z., X.T., R.X., Z.H., Y.S., and C.H. primarily performed the experiments. Q.S., X.Q., S.H., H.Z., Y.S., Y.L., X.Z., H.J., X.T., R.X., Z.H., C.H., and X.L. contributed to the data analyses and discussion. Q.S., Y.L., X.Z., X.Q., and H.J. prepared the figures. Q.S., Y.L., X.Q., H.Z., and S.H. wrote the manuscript with contributions from other authors. Peer review Peer review information Nature Communications thanks Diana Nelli and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available. Funding We acknowledge the support from the Tsinghua Shenzhen International Graduate School-Shenzhen Pengrui Young Faculty Program of Shenzhen Pengrui Foundation (No. SZPR2023001, S.H.) and Shanghai Pujiang Program (No. 22PJD025, Q.S.). This work was also supported by National Natural Science Foundation of China (No. 22020102003 (H.Z.), 52302168 (X.Q.), 62288102 (X.Q.), 22371162 (S.H.), 22575125 (X.L.)), the Guangdong Innovative and Entrepreneurial Research Team Program (2023ZT10C040, S.H.), the National Key Research and Development Program of China (No. 2023YFF0714200, S.H.), Shenzhen Medical Research Fund (No. B2502021 and B2302021, S.H.) and Guangdong Basic and Applied Basic Research Fund (No. 2024A1515010713, S.H.), and the Overseas Research Cooperation Fund of Tsinghua SIGS (No. HW2023002, S.H.). Data availability All data supporting the findings of this study are available within the article. The source data that support the findings of the study are available from the Figshare at (10.6084/m9.figshare.32227674). Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Xian Qin, Email: ifexqin@fjnu.edu.cn. Sanyang Han, Email: hansanyang@sz.tsinghua.edu.cn. Hongjie Zhang, Email: hjzhang2019@mail.tsinghua.edu.cn. Qianqian Su, Email: chmsqq@shu.edu.cn. Supplementary information The online version contains supplementary material available at 10.1038/s41467-026-74447-5. References Associated Data Supplementary Materials Supplementary Information Transparent Peer Review file Data Availability Statement All data supporting the findings of this study are available within the article. The source data that support the findings of the study are available from the Figshare at (10.6084/m9.figshare.32227674).