General phase segregation and phase pinning effects in lanthanide-doped lead halide perovskite with dual-wavelength lasing Authors: Junyu He, Jun Luo, Biyuan Zheng, Feng Jiang, Tongqing Sun, Yaonan Xiong, Zhou Li, Jing Wu, Chenglin He, Jiahao Liu, Tingzhao Fu, Mengjian Zhu, Shulin Chen, Honglai Li, Mengqiu Cai, Zhihong Zhu, Fang Wang, Weihao Zheng, Xiujuan Zhuang Published: 2025 ABSTRACT AbstractA strategy of lanthanide-ion doping into dual-halogen-alloyed perovskites CsPb(XxY1-x)3 (X, Y = Cl, Br, I) via chemical vapor deposition is introduced, obtaining a series of high-quality, stable microplates. Under continuous light excitation, each sample exhibits highly stable dual-band photoluminescence emission, whereby pairwise combinations of the three halogens enable photoluminescence to cover the red, green, and blue spectral regions. Corresponding high-performance dual-wavelength lasers are achieved. The segregated phase domains are tens of nanometers in size with well-defined boundaries. Theoretical calculations indicate that lanthanide-ion doping promotes phase segregation and facilitates ion migration in the alloyed case, while suppressing it in the phase-segregated state, producing a phase-pinning effect. This mechanism imposes opposite trends on the migration barrier in alloyed versus deployed domains, simultaneously driving halide segregation and pinning ion migration in segregated phases. Our work simultaneously enhances stability and broadens the bandgap-engineering for lead-halide perovskites, accelerating their entry into next-generation optoelectronics.Subject terms: Semiconductor lasers, Structural properties Introduction Lead halide perovskites (LHPs) exhibit outstanding optoelectronic properties, including high absorption coefficients, high photoluminescence (PL) quantum yields, remarkable defect tolerance, tunable bandgaps, and excellent charge transport characteristics1–4. These distinctive features have enabled their widespread application in various optical fields, including light-emitting diodes (LEDs)5–8, photodetectors9–11, visual displays12–14, and lasers15–17. However, the soft lattice nature of LHPs makes them highly susceptible to rapid halide ion migration when exposed to external stimulation, such as applied electric fields, light irradiation, and temperature changes18–22. Such ion migration can induce interfacial halide ion diffusion or segregation, electrode polarization, and hysteresis phenomena23–25. These issues not only undermine the stability of the devices but also induce substantial performance degradation. Moreover, ion migration can erase the hetero-interfaces, which present a significant obstacle to constructing stable multi-band composite structures based on LHPs. Suppressing ion migration to obtain perovskites with both stability and multi-band composite structures remains a significant challenge that urgently needs to be addressed, as it is essential for advancing perovskite bandgap engineering and unlocking novel functional applications. Recently, numerous studies have focused on suppressing ion migration in perovskite materials. Chang et al. have significantly suppressed ion migration in perovskite solar cells by generating a highly ordered two-dimensional (2D) perovskite on the surface of a 3D perovskite film using a meta-amidinopyridine ligand and a solvent post-dripping step, thereby enhancing the performance and stability of the devices26. Wang et al. effectively suppressed ion migration by introducing thioacetylacetamide hydrochloride to bind with the cations and anions (named halide locking) in the perovskite, thereby reducing the formation of ion migration pathways27. Interestingly, Hong et al. proposed a phase pinning strategy, which stabilizes the disordered phase by suppressing the ordering transition of the spacer cations in two-dimensional perovskites through the introduction of a small amount of pentylammonium or butylammonium as dopants28. However, these ion-migration-suppression strategies all rely on the doping of organic ligands or low-valence cations. While these dopants can transiently immobilize ionic pathways, they readily undergo hydrolysis or lattice reconstruction under humid-heat conditions, compromising long-term stability29,30. More importantly, they focus solely on suppressing migration to extend device performance, without probing the design and application potential of tailored band structures. On the other hand, phase segregation is a common occurrence in mixed-halide perovskite alloys when subjected to light, heat, or electrical excitation31. This process, driven by the migration of halide ions, results in the enrichment of different halide elements and the formation of multi-band composite structures. Regrettably, in the absence of external stimulation, spontaneous ion migration and diffusion typically dismantle these multi-band composite structures, reverting them to single-band alloyed configurations. However, if ion migration suppression, i.e., phase pinning, can be achieved based on phase segregation, the segregated pure halide phase can be preserved even after the removal of external stimuli, thereby realizing their stability. This structure will fully leverage the suppression of ion migration to re-engineer perovskites, stable multi-band composite structures fundamentally. In this work, by doping with various Ln (lanthanide) ions (Er3+, Yb3+, Nd3+, and Eu3+) through a simple one-step chemical vapor deposition (CVD) strategy, we fabricate Ln-CsPb(XxY1-x)3 (X, Y = Cl, Br, I) microplates (MPs) with intrinsic phase segregation. Representative Er-CsPb(XxY1-x)3 MPs are selected for a comprehensive and in-depth investigation from multiple aspects, including micro-structure, optical properties, carrier dynamics, and theoretical calculations on ion migrations. Significantly, the halide ion migration in these MPs is effectively suppressed, therefore stabilizing these segregated phases and thus achieving phase pinning. Based on the simultaneous phase segregation and phase pinning, we accomplish the engineered multi-band composite structures with high stability. Stable blue and green dual-band PL emissions in Er-CsPb(ClxBr1-x)3 are observed, as well as green and red, blue and red dual-band PL emissions for Er-CsPb(BrxI1-x)3 and Er-CsPb(ClxI1-x)3 MPs, respectively. Utilizing these MPs with stable dual-band emissions, we achieve three types of dual-wavelength lasers with relatively low thresholds and short lifetimes. Such good performances of lasing indicate the excellent optical features and high quality of the lattice for cavities in these MPs. Theoretical calculations clarify the phase segregation and pinning mechanism in Ln-doped LHPs. Our findings provide a general and straightforward strategy for fabricating mixed halide LHPs with both phase segregation and phase pinning, offering a practical approach to enhancing the stability of perovskites and significantly expand their potential for future optoelectronic applications. Results and discussion Optical properties The Er-CsPb(XxY1-x)3 MPs were prepared using a precisely controlled CVD method (see Methods and Supplementary Fig. 1). Since the source materials contain two species of halogen elements and a small amount of erbium halide, all-inorganic mixed halogen perovskite with three-dimensional structures doped with Er3+ were synthesized, as depicted in Fig. 1a. When two different halide ions coexist in the perovskites, they typically appear in a homogeneous alloy state, resulting in a single bandgap between that of two pure species, which is a classical method of tuning the bandgap of a compound semiconductor32. Therefore, only a single band-edge recombination PL can usually be observed in mixed halogen perovskites33. However, quite an unusual phenomenon was observed in our case. Fig. 1b shows the PL spectra and emission images (insets in Fig. 1b) of the Er-CsPb(ClxBr1-x)3, Er-CsPb(BrxI1-x)3, and Er-CsPb(ClxI1-x)3 MPs, where each sample exhibits dual-color emission bands responsible for the dual peaks of the PL. The Er-CsPb(ClxBr1-x)3 shows a blue band at 426.1 nm and a green band at 518.6 nm, corresponding to the Cl-rich and Br-rich phase, respectively. Similarly, Er-CsPb(BrxI1-x)3 shows a green band (Br-rich phase) and a red band (I-rich phase) at 538.3 and 694.3 nm, and the Er-CsPb(ClxI1-x)3 shows a blue band (Cl-rich phase) and a red band (I-rich phase) at 431.8 and 689.7 nm, respectively. Notably, the commonly observed fluorescence of Er3+ at 525 nm and 550 nm corresponding to the 2H11/2 → 4I15/2 and 4S3/2 → 4I15/2 transitions, disappeared in our samples, because some f-f transitions are forbidden when lanthanide ions are located in the centers of a highly symmetrical lattice (see Supplementary Note 1 for details)34. For example, the 2H11/2 → 4I15/2 transition is a type of hypersensitive transition susceptible to the lattice environment. The spectral intensity of such hypersensitive transitions depends on the local symmetry of the crystal field surroundings in which the rare-earth ions are located. Specifically, when the rare earth ion is located at the symmetric center of the crystal, this type of transition is forbidden (see details in Supplementary Fig. 2). Our materials are in well-defined morphology with a single-crystal structure, where the Er3+ ions are likely predominantly incorporated at the B-site, which is the symmetric center of the PbX6 octahedra. The above analysis is consistent with our material properties and previous theoretical calculations35. Optical characterization of Er-CsPb(XxY1-x)3 MP. a Lattice structure diagram of Er-CsPb(XxY1-x)3 MPs. b PL spectra of Er-CsPb(ClxBr1-x)3, Er-CsPb(BrxI1-x)3 and Er-CsPb(ClxI1-x)3 MPs, respectively. Scale bar: 2 μm. The insets are the corresponding PL images. c Corresponding emission intensities of each phase in three MPs as a function of the exciting laser shots. d PL spectra of Er-CsPb(ClxBr1-x)3, Er-CsPb(BrxI1-x)3 and Er-CsPb(ClxI1-x)3 MPs after lasing. The insets are the corresponding images during lasing. Scale bar: 2 μm. This Er-doping-induced stable phase segregation fundamentally differs from the photo-, electro-, or thermal-induced phase segregation in undoped LHPs, which has been widely reported previously36. The phase segregation observed in our case is not affected by external excitation, including a certain degree of thermal and photoexcitation. As shown in Fig. 1c, these dual-band emissions exhibit total stability for their intensities with increasing the number of laser shots. In addition, the emissions of Er-CsPb(ClxBr1-x)3, Er-CsPb(BrxI1-x)3, and Er-CsPb(ClxI1-x)3 MPs still maintain thermal stability at 600, 450, and 450 K, respectively, but experience failure at 650 (Supplementary Fig. 3), 500 (Supplementary Fig. 4), and 500 K (Supplementary Fig. 5). These significant dual-band emission properties and their relevant stability differ from those of the undoped CsPb(XxY1-x)3 MPs, which only exhibit a single alloying peak at the initial light excitation time (Supplementary Fig. 6). Temperature-dependent PL decay measurements under a fixed bias, analyzed within the Arrhenius framework, reveal the activation energies for halide ion migration in Er-CsPb(ClxBr1-x)3, Er-CsPb(BrxI1-x)3, and Er-CsPb(ClxI1-x)3 MPs. As shown in Supplementary Figs. 7-10, the calculated barriers are 0.99 eV (Cl⁻) and 1.08 eV (Br⁻), 1.19 eV (Br⁻) and 1.07 eV (I⁻), 1.07 eV (Cl⁻) and 1.16 eV (I⁻), respectively. The values that substantially exceed the previously reported range of 0.20-0.65 eV37–40. Based on the above results, we believe that Er3+ doping can stabilize and pin different segregated halide phases, thereby hindering halide ion migration in the LHPs. Even more interesting, lasing can be conveniently implemented at each PL band under an optical pump. Each MP is a dual-wavelength semiconductor laser (Fig. 1d). Obvious coherent fringes can be observed among these MPs during lasing (insets in Fig. 1d). Dual-wavelength lasing Confirmatory laser experiments, including laser threshold effect, spectral broadening, and carrier lifetimes, were performed to confirm the above dual-wavelength laser. PL spectra of the representative Er-CsPb(ClxBr1-x)3 MP under several selected pump fluences are shown in Fig. 2a. At low pump fluence, such as 10.1 μJ cm-2, the spectrum exhibits two PL peaks with a full width at half maximum (FWHM) value of 16 ~ 23 nm. As the pump fluence increases, the sharp lasing peak first appears at the Br-rich phase. When the pump fluence is increased to 29.0 μJ cm-2, the Cl-rich phase begins lasing and eventually displays dual-wavelength lasing behavior under a pump fluence of 64.8 μJ cm-2. At this time, the FWHM values were narrowed to 1.3 ~ 1.8 nm. Fig. 2b plots the PL intensity as a function of the pump fluence, indicating classic S-shape curves. Figure 2c, d presents their corresponding FWHM values as a function of the pump fluence, respectively. The lasing threshold pump fluences (Pth) of 15.4 and 29.0 μJ cm-2 are obtained for the Cl-rich and Br-rich phases, respectively. Time-resolved spectroscopy is conducted to investigate the lasing lifetime performance. Fig. 2e and f respectively give the streak camera images of the Cl-rich and Br-rich phases from the Er-CsPb(ClxBr1-x)3 MP when lasing. The bright dots in the two images imply the short lasing lifetime when the pump fluence is 1.2 Pth. After the single exponential fit with deconvolution of the instrument response function (IRF), time constants of 3.2 and 2.4 ps are determined for the Cl-rich phase and Br-rich phase, respectively. Similar dual-wavelength lasing behavior can also be observed in the Er-CsPb(BrxI1-x)3 (Supplementary Figs. 11 and 12) and Er-CsPb(ClxI1-x)3 MPs (Supplementary Figs. 13 and 14). Zoomed-in dual-wavelength lasing spectra for Er-CsPb(ClxBr1-x)3, Er-CsPb(BrxI1-x)3, and Er-CsPb(ClxI1-x)3 MPs are shown in Supplementary Fig. 15, and the cavity-mode dynamics underlying the lasing behavior are presented in detail in Supplementary Note 2. Lasing behavior characterization of Er-CsPb(ClxBr1-x)3 MP. a Representative lasing evolution of the Er-CsPb(ClxBr1-x)3 MP under several selected pump fluences. b PL intensities as a function of the pump fluence for both Cl-rich and Br-rich phases. Corresponding FWHM statistics of c, Cl-rich phase and d, Br-rich phase as a function of the pump fluence. e Streak camera image of Cl-rich after lasing and the corresponding dynamic curve. f Streak camera image of Br-rich phase after lasing and the corresponding dynamic curve. IRF is indicated. Energy transfer dynamics Time-resolved PL (TRPL) spectroscopy is also employed to investigate the carrier dynamics of the Er-CsPb(XxY1-x)3 MPs before lasing. Fig. 3a and b give the streak camera images collected from the PL of the Cl-rich and Br-rich phases of a typical Er-CsPb(ClxBr1-x)3 MP, respectively. The PL decay curves for both phases are extracted from the areas between the white-dashed lines and are shown in Fig. 3c. These dynamic curves are fitted by a multi-exponential function with deconvolution of the IRF (see Supplementary Note 3 for details)41. The fitting process and results include PL decay time constants, and the corresponding amplitudes are listed in Supplementary Table 1. Three pristine perovskite samples with pure phases, CsPbCl3, CsPbBr3, and CsPbI3, are used to compare and study the carrier dynamics (Supplementary Fig. 16 and Supplementary Table 2). Typically, pristine CsPbX3 without Er3+ doping exhibits time constants with both positive amplitudes, which are attributed to carrier recombination through surface states with a shorter time constant and intrinsic states with a longer time constant, respectively42,43. However, the PL decay of the Cl-rich phase in the Er-CsPb(ClxBr1-x)3 MP exhibits three components, all with positive amplitudes, and the rise time of the dynamic curve is close to the IRF. In contrast, the dynamic curve of the Br-rich phase is much longer than the IRF and exhibits a longer lifetime than that in the Cl-rich phase. The Br-rich phase has three time constants with one negative amplitude and two positive values. The negative amplitude with a prolonged rise edge of the dynamic curve typically implies an extra significant carrier import into the Br-rich phase44. The above results demonstrate that a band-structure realignment occurs, indicating a robust energy transfer process between the Cl-rich phase and the Br-rich phase. Carrier dynamics of the representative Er-CsPb(ClxBr1-x)3 MP. a Streak camera images collected in the Cl-rich phase emission band of Er-CsPb(ClxBr1-x)3 MP. b Streak camera images collected in the Br-rich phase emission band of Er-CsPb(ClxBr1-x)3 MP. c PL decay curves of Cl-rich phase and Br-rich phase perovskites of Er-CsPb(ClxBr1-x)3 MP. The light red and black shaded regions mark the curve risetimes of the Cl-rich and Br-rich phases, respectively. d Schematic diagram of the charge funneling and energy/charge transfer mechanism in the Er-CsPb(ClxBr1-x)3 system. Specifically, previous studies indicate that CsPbCl3 and CsPbBr3 usually adopt a type-I band alignment45,46. In this band configuration, both electrons and holes cross the interface within sub-picoseconds to a few picoseconds, enabling ultrafast charge transfer. Concurrently, dipole-assisted Förster resonance energy transfer (FRET) may also occur. With a donor lifetime of several nanoseconds and a Förster radius of a few nanometers, sub-picosecond transfer is possible if the donor-acceptor interface is atomically sharp47–49. Both routes are plausible. Which one is dominant is dependent on the exact interface width. Nevertheless, given the possible occurrence of both FRET and charge transfer, the slowest transfer at the interface does not exceed a few picoseconds, much faster than the experimentally observed timescale of hundreds of picoseconds. On the other hand, the carrier population at the boundary is rapidly depleted by highly efficient interfacial energy transfer, resulting in a steep concentration gradient. Subsequent carrier diffusion toward the interface proceeds at the diffusion-limited rate and is therefore slower. The measured time constants of hundreds of picoseconds naturally follow from convolving this funneling step with the preceding ultrafast direct transfer. Similarly, the carrier funneling phenomenon was also observed in Er-CsPb(BrxI1-x)3 (Supplementary Fig. 17 and Supplementary Table 3) and Er-CsPb(ClxI1-x)3 MPs. (Supplementary Fig. 18 and Supplementary Table 4) MPs. The TRPL confirms the phase segregation phenomenon from the bandgap realignment caused by the mixed halide dealloying. In addition, we selectively pumped the Br-rich domains of Er-CsPb(ClxBr1-x)3 MP by using 470 nm pulsed laser excitation and monitored their PL decay (Supplementary Fig. 19 and Supplementary Table 5). The decay curve shows no rise component and is accurately fitted by a bi-exponential decay, further providing direct evidence for the existence of energy transfer in doping MPs. Furthermore, power-dependent PL experiments of Er-CsPb(ClxBr1-x)3 reveal that the integrated PL intensity ratio of the Cl-rich to the Br-rich phase increases as the excitation power is increased (Supplementary Fig. 20). This observation is attributed to the elevated density of photo-generated carriers, which intensifies nonradiative recombination events during charge funneling and thereby lowers the overall energy transfer efficiency. The relevant analysis is detailed in Supplementary Note 4, based on the results (Supplementary Fig. 21) obtained from pump-fluence-dependent TRPL experiments. Consequently, the relative emission of the Br-rich phase (acceptor) is suppressed. An analogous trend is observed in both Er-CsPb(BrxI1-x)3 and Er-CsPb(ClxI1-x)3, as shown in Supplementary Fig. 22 and Supplementary Fig. 23, respectively. Morphological and structural analyses According to the emission peak position, the specific values of the components in each phase can be defined through the conventional bowing equation (Supplementary Note 5)50: Eg1 and Eg2 are the bandgaps of the corresponding single-component halide perovskite, and Eg(x) is bandgap of mixed-halide perovskites with component values of x. m is the quadratic term coefficient. The Er-CsPb(ClxBr1-x)3 is a mixed-phase MP consisting of a Cl-rich phase with a chemical formula of Er-CsPb(Cl0.95Br0.05)3 and a Br-rich phase with a chemical formula of Er-CsPb(Cl0.09Br0.91)3. For Er-CsPb(BrxI1-x)3, the Br-rich and I-rich phase is Er-CsPb(Br0.93I0.07)3 and Er-CsPb(Br0.10I0.90)3, and for Er-CsPb(ClxI1-x)3, its Cl-rich and I-rich phase are Er-CsPb(Cl0.94I0.06)3 and Er-CsPb(Cl0.04I0.96)3. According to the calculated results, the above blue, green, and red bands are close to the band-edge PL of pristine pure phase perovskite of CsPbCl3, CsPbBr3, and CsPbI3, respectively (Supplementary Fig. 24). Why does the usually alloyed system of the mixed halide perovskites exhibit a significant phase segregation performance? Materials characterization was first performed to clarify the mechanism. From the corresponding optical images of the Er-CsPb(ClxBr1-x)3, Er-CsPb(BrxI1-x)3, and Er-CsPb(ClxI1-x)3 MPs (Fig. 4a-c), it can be seen that all three MPs feature a regular and crystalline morphology, showing the equilateral square shape with the side lengths of several micrometers. From the scanning electron microscopy (SEM) image shown in Fig. 4d, most samples exhibit good shape, confirming the preferable homogeneity during the growth process. Fig. 4e presents the high-magnification SEM image for a representative single Er-CsPb(ClxBr1-x)3 MP, which shows a well-defined morphology without obvious surface defects. Corresponding energy dispersive X-ray (EDX) for element mapping images are shown in Fig. 4f-j, respectively. The results indicate that the Er elements accounted for 3.1% of the total element content in Er-CsPb(ClxBr1-x)3 MP. Supplementary Fig. 25 provides similar results for Er-CsPb(BrxI1-x)3 and Er-CsPb(ClxI1-x)3 MPs, where the Er element concentrations are 3.3% and 4.4% of the total element content, respectively. The atomic ratio of Cs:(Pb+Er):(Cl+Br) is about 1:0.93:3.05, close to the quantitative ratio of CsPbX3 perovskite, which is consistent with the study that the Er3+ doped into the lattice are more accessible to replace the B-site ions (Pb2+)35. Based on the EDX results, the overall chemical formula of the three selected MPs can be given as Er-CsPb(Cl0.83Br0.17)3, Er-CsPb(Br0.81I0.19)3, and Er-CsPb(Cl0.83I0.17)3, respectively. Furthermore, X-ray photoelectron spectroscopy (XPS) measurements on Er-CsPb(Cl0.83Br0.17)3 (Supplementary Fig. 26), Er-CsPb(Br0.81I0.19)3 (Supplementary Fig. 27), and Er-CsPb(Cl0.83I0.17)3 (Supplementary Fig. 28) MPs reveal consistent elemental stoichiometries and Er-doping concentrations, demonstrating their compositional homogeneity. Morphology and structure characterization of Er-CsPb(XxY1-x)3 MP. a-c The optical images of Er-CsPb(ClxBr1-x)3, Er-CsPb(BrxI1-x)3 and Er-CsPb(ClxI1-x)3, respectively. Scale bar: 2 μm. d Large-area low-magnification SEM image of typical Er-CsPb(ClxBr1-x)3 MPs. Scale bar: 50 μm. e High-magnification SEM image of a selected Er-CsPb(ClxBr1-x)3 MP. Scale bar: 2 μm. f–j Element mapping of the typical Er-CsPb(Cl0.83Br0.17)3 MP. k XRD patterns of CsPb(Cl0.83Br0.17)3 and Er-CsPb(Cl0.83Br0.17)3 MPs. l Lattice structure diagram before and after Er3+ doping. m Hyperspectral CL mapping acquired within the wavelength range of 360-440 nm and n 460-560 nm. Scale bar: 50 nm. o CL acquired at positions 1 (Cl-rich phase) and 2 (Br-rich phase) marked in panels i and j. Fig. 4k shows the X-ray diffraction (XRD) patterns of the Er-CsPb(Cl0.83Br0.17)3 MPs compared with the undoped CsPb(Cl0.83Br0.17)3 MPs. Two kinds of MPs possess diffraction peaks that match the (100), (110), and (200) crystal planes of the cubic CsPbX3 lattice structure, which indicates that Er3+ doping does not change the lattice structure of the perovskite. Compared to the undoped sample, the prominent diffraction peak of Er-CsPb(Cl0.83Br0.17)3 shifted to a larger angle, apparently. According to Bragg’s law, a lattice strain of ε = −0.06 is calculated (Fig. 4l, see Supplementary Note 6). This value is markedly larger than the ε = −0.023 to −0.005 reported for analogous systems51–54, underscoring that Er3+ doping in our case induces a substantial lattice shrinkage. This pronounced lattice contraction probably arises from two synergistic origins. Firstly, the substantial ionic radius mismatch (0.89 Å for Er3+ vs. 1.19 Å for Pb2+) imposes an immediate and sizable shrinkage, with strain magnitude scaling directly with the size disparity. Secondly, the aliovalent substitution requires charge compensation through Pb vacancies or interstitial halides. These defects further distort the lattice, and the local collapse surrounding each Pb vacancy amplifies the strain field to an exceptional degree. Besides, the high concentration doping of Er3+ could induce more pronounced lattice distortions. Both the EDX and the XRD results have proved that Er3+ is successfully doped into the lattice of the host perovskite. SEM-based cathodoluminescence (SEM-CL) combines nanometer spatial resolution with spectroscopic discrimination, offering an ideal tool for resolving the intrinsic microstructure of these MPs. Hyperspectral CL mapping reveals a pronounced spatial anticorrelation between the blue and green domains (Fig. 4m and n), with individual CL spectra extracted from each domain exhibiting single-peak emission at 407 and 522 nm, respectively (Fig. 4o). The blue and green domains correspond to the Cl-rich and Br-rich phases, respectively. The irregular domain size of segregated phases is approximately tens of nanometers, and the domain interfaces are sharp. Their morphology likely extends from the two-dimensional surface into the three-dimensional bulk. Compared to the PL of each segregated phase, the CL bands are noticeably blue-shifted and broadened, reflecting the non-equilibrium carrier distribution created by the high-energy electron beam, local lattice heating, and surface-mediated phenomena intrinsic to CL excitation. The SEM-CL results strongly demonstrate the phase separation and phase pinning phenomena in these materials. This intrinsic phase segregation and phase pinning are also observed in Yb3+, Nd3+, and Eu3+ doped systems, as shown in Supplementary Fig. 29. Except for some fluorescence transition from Eu3+ is observed, the fluorescence from the Yb3+ and Nd3+ ions are completely disappeared, because of the restriction on the f-f transition rules governing within the perovskite lattice (Supplementary Fig. 2 and Supplementary Note 1). The pronounced shift of diffraction peaks in the XRD patterns (Supplementary Fig. 30), corroborated by the homogeneous distribution of rare earth elements on a larger scale, as revealed through SEM-EDX mapping (Supplementary Figs. 31-33, b-f), indicates successful lattice doping. Nevertheless, the phase segregation induced by these alternative Ln3+ ions remains conspicuously incomplete compared with the near-complete separation achieved by Er3+ doping. Consequently, the nominally Cl-rich phase retains appreciable alloying, manifesting as a progressive shift of the blue emission band. Moreover, these MPs exhibit compromised lattice quality and irregular morphologies as evident in the SEM images for Yb3+-doped (Supplementary Fig. 31), Nd3+-doped (Supplementary Fig. 32), and Eu3+-doped (Supplementary Fig. 33) MPs, respectively. These unsatisfactory morphologies ultimately preclude the observation of lasing or even amplified spontaneous emission in these MPs. While the doping efficacy of these alternative Ln3+ ions does not match the performance exhibited by Er3+, the underlying phenomena remain qualitatively identical. This convergence reveals that the heterovalent substitution of Pb2+, along with the consequent phase segregation and pinning, operates through a standard and broadly applicable mechanism. Theoretical calculation To investigate the specific mechanism by which Er3+ doping promotes phase segregation and phase pinning, we performed first-principles calculations and analyzed the effect of Er3+ from the perspective of formation energy and ion migration. To clarify the types of defects that doping may induce, we first calculate the formation energies of different defect types in Er-CsPb(ClxBr1-x)3. According to Supplementary Fig. 34, only the defect in which Er3+ substitutes Pb2+ (PbEr) exhibits a negative formation energy, indicating that this defect type is the most likely to form. Therefore, when calculating the formation energy of each phase and the ion migration barrier in them, we only considered this type of defect. The CsPb(Cl0.83Br0.17)3 with a typical Cl/Br ratio was selected as the analytical object for theoretical calculations. Two types of lattice structural configurations, the homogeneous alloy and the regional halogen aggregation state, were calculated and listed for both Er-doped and undoped CsPb(Cl0.83Br0.17)3 in Supplementary Table 6. For the Er-doped material, the formation energy of the phase segregated case will reduce by 0.97 eV compared to its homogeneous alloy case (Fig. 5a), while for the materials without Er-doping, the formation energy of the phase segregated case will rise by 0.10 eV compared to their alloy state (Fig. 5b). From the perspective of thermodynamics, the chemical reaction with Er-doping will move to the direct of forming Br-rich and Cl-rich phase, and the phase segregation state will be much more stable than that in alloy state, which is precisely opposite in the case without doping. The mechanism of Er-doping facilitates phase segregation, which can be understood as follows: when one halogen species aggregates, the local chemical potential and electronegativity in this region change significantly. These variations will lead to a more significant charge density gradient, which makes the interaction between ions uneven and further intensifies the chemical instability of adjacent regions, prompting the aggregates region to separate from the other areas and form pinning phase segregation. Theoretical calculation of the formation energy and ion migration barrier. a Formation energy of Er-CsPb(Cl0.83Br0.17)3 in homogeneous alloy case and regional halogen aggregation case, where the regional halogen aggregation case will be more stable. b Formation energy of CsPb(Cl0.83Br0.17)3 without doping in homogeneous alloy case and regional halogen aggregation case, where the alloy phase will be more stable. c Theoretical calculations of intra- octahedral migration barriers for Cl⁻ in CsPb(Cl0.83Br0.17)3, Er-CsPb(Cl0.83Br0.17)3, and Er-CsPbCl3. d Schematic diagram of Er3+ doping induced phase segregation and phase pinning. The ion migration barrier represents a fundamental criterion for assessing the migration propensity of halogen ions. Since the ratio of Cl/Br in CsPb(Cl0.83Br0.17)3 is about 5:1, the migration capacity of the dominant Cl⁻ can reflect the performance of the primary material components. Based on this, we further calculated the intra- and inter-octahedral migration barriers of Cl⁻ in undoped alloy CsPb(Cl0.83Br0.17)3, Er-doped alloy Er-CsPb(Cl0.83Br0.17)3, and Er-doped single halide CsPbCl3 (Er-CsPbCl3) (Supplementary Table 7). The Er-CsPbCl3 can be regarded as one of the phases that form after a phase segregation process. As shown in Fig. 5c, the intra-octahedral migration barrier of Cl⁻ decreases from 0.56 eV in CsPb(Cl0.83Br0.17)3 to 0.51 eV in Er-CsPb(Cl0.83Br0.17)3. In contrast, the halide ion migration barrier will significantly increase from 0.51 eV in Er-CsPb(Cl0.83Br0.17)3 to 1.58 eV in one of the deallyed phases, i.e. Er-CsPbCl3. The calculation results for the inter-octahedral migration exhibit a similar trend (Supplementary Fig. 35). Combined with the formation energy results, the Er-doping-induced phase segregation can be decomposed into two independent formation processes from a perspective of transition state theory. As shown in Fig. 5d, the alloy phase CsPb(Cl0.83Br0.17)3 without doping and the Er-doped CsPbCl3 or CsPbBr3 can be regarded as the initial state and the final states, respectively. When Er3+ is introduced into the CsPb(Cl0.83Br0.17)3, a corresponding alloy phase Er-CsPb(Cl0.83Br0.17)3 may initially form. However, this state is thermodynamically unstable and can be regarded as a transition state. Furthermore, this state exhibits a reduced ion migration barrier of 0.51 eV. At the subsequent process, the system may evolve into a dealloyed state with much lower formation energy, as presented in Fig. 5a. Once the halide ion aggregation occurs and the dealloyed phase (Er-CsPbCl3 or Er-CsPbBr3) is formed, ion migration becomes considerably more difficult in them because of their significantly increased barrier. This is similar to halogen ion aggregation and dropping into a potential well, where they are bound within it, which suggests that the halide ions have been effectively pinned. Overall, Er3+ facilitates ion migration in the alloy phase, while it plays a suppressive role in halide ion migration in each dealloyed phase. These results demonstrate that Er3+ doping not only facilitates phase segregation to generate multi-band tunability but also effectively suppresses ion migration through a phase-pinning phenomenon to enhance the overall stability of the system (Fig. 5d). It is precisely because the different influences in alloy and dealloy states on the potential barrier exhibit opposite trends that our phase separation material achieves a stable final state. In summary, an strategy that simultaneously utilizes phase segregation and phase pinning effects to achieve stable multi-band composite structures in Er-CsPb(XxY1-x)3 MPs is demonstrated. Through a simple one-pot CVD method, three stable multi-band composite structures were fabricated, which exhibit three types of dual-band PL emission spanning the red, green, and blue spectral regions with stable photoexcitation stability. Three types of dual-wavelength lasers have been realized, featuring thresholds in the tens of μJ cm-2 and lifetimes on the order of several picoseconds. These high-quality dual-wavelength lasers, in turn, demonstrate that the prepared multi-band composite structures possess high-quality lattice structures and high stability. TRPL spectra experiments manifest a robust energy transfer within these separated phases. From the perspective of forming energy and ion migration through first-principles calculation, lanthanum ion doping makes the local halogen aggregation more stable than the homogeneously mixed halogen distribution case, which is conducive to local phase segregation and phase pinning. This study develops a strategy for controlling halide ion migration and constructing multi-band composite structures with high stability and lattice quality for LHPs. This advancement significantly broadens the application scope of LHPs in photonics and optoelectronics. Methods Sample Preparation. Er-CsPb(XxY1-x)3 MPs were grown using a CVD method. For Er-CsPb(ClxBr1-x)3 MPs, ErCl3 (99.999%, Alfa), CsBr (99.999%, Alfa), PbCl2 (99.999%, Alfa) powders are mixed in a clean alumina boat with a mass ratio of 3:1:1 and then placed in the heating zone of a tubular furnace. For Er-CsPb(BrxI1-x)3 MPs, ErBr3 (99.999%, Alfa), CsI (99.999%, Alfa), PbBr2 (99.999%, Alfa) powers are mixed in a clean alumina boat with a mass ratio of 6:1:3 and then placed in the heating zone of a tubular furnace. For Er-CsPb(ClxI1-x)3 MPs, ErCl3 (99.999%, Alfa), CsI (99.999%, Alfa), PbCl2 (99.999%, Alfa) powers are mixed in a clean alumina boat with a mass ratio of 3:1:1 and then placed in the heating zone of a tubular furnace. The freshly cleaved fluorophlogopite mica is positioned downstream of the furnace as the substrate. To clean the impurity gas, a high flow rate of carrier gas Ar is maintained for 10 minutes (>500 sccm). Then, the pressure of the system is decreased to less than 20 mTorr by a vacuum pump and a constant flow of ~20 sccm of carrier gas during the entire growth. The corresponding thermodynamic parameters were set according to the different target products: for Er-CsPb(ClxBr1-x)3, the temperature was raised to 820 °C and held for 6 minutes; For Er-CsPb(BrxI1-x)3, the temperature was raised to 960 °C and held for 14 minutes; For Er-CsPb(ClxI1-x)3, the temperature is heated to 830 °C and held for 6 minutes. The entire system is naturally cooled down after the growth process. Optical Measurements. Steady-state PL spectra and TRPL measurements for the Er3+ doping samples were detected based on a confocal microscope system (WITec, alpha-300). A mode-locked Ti: sapphire laser (Tsunami) at 800 nm (pulse width 80 fs, repetition frequency 80 MHz) is introduced to a BBO crystal to obtain a second harmonic generation with 400 nm wavelength as the excitation light source. Samples doped with other lanthanide ions were excited by a 385 nm pulsed laser with the same system. Dark-field image measurements were conducted by introducing the corresponding laser through the 50-fold lens to excite the MPs without any other illumination. For lasing experiments, the same confocal microscope was employed. The 800 nm laser was amplified by a regenerative amplifier (Spitfire Ace 100, 1 kHz) and passed through a BBO crystal to produce 400 nm light (1 kHz) for excitation. An Olympus 20× objective lens focused the lasing beam to create a uniform spot with a diameter of 200 μm on the sample. A spectrometer detected the PL and lasing signals. TRPL experiments were measured by reflecting the light signal into the streak camera (C10910, Hamamatsu) by Ag mirrors. Theoretical Calculations. Density functional theory (DFT) calculations are implemented by the Vienna ab initio Simulation Package (VASP). In this study, the projected augmented wave (PAW) method was used to describe the interaction between valence electrons and ionic nuclei. The energy cutoff of the plane wave base group is set to 300 eV. The convergence standard of the force is 0.02 eVÅ-1. All calculations of the defect structure, i.e., formation energy, are performed in 3×3×3 supercells, so the K-point grid is set to 1×1×1. Material formation energy Ef refers to the change in energy that introduces atoms from the standard state of the element into the material under study. In the calculation, the formation of the material can be calculated by the following formula: Where Etotal is the total energy of the material, Eielement is the energy of the ith element in its standard state, and ni is the number of moles of the ith component in the material. The total energy Etotal of the material is calculated by structural optimization, and the energy of each element in its standard state is taken from its reference structure. A migration barrier Eb, is a barrier that describes the energy required for an ion or atom to move from one location to another, which can be calculated using the following formula: Where ETS is the energy of the transition state, and Estart and Eend are the energies of the initial state (starting positions) and final state (target positions), respectively. The energy of the transition state is searched by linearly inserting multiple intermediate configurations between the starting and target locations. Both intra- and inter-octahedral Cl⁻ migration pathways were examined within a 3×3×3 CsPb(Cl0.827Br0.173)3 supercell measuring ~17.6 Å on each side and containing 135 atoms (27 ABX3 units). This size guarantees that the closest distance between any migrating defect and its periodic image exceeds 10 Å, eliminating spurious image interactions. For the intra-octahedral process, a single octahedron embedded in the 3 × 3 × 3 matrix was targeted; for the inter-octahedral process, two neighboring octahedra within the same supercell were considered. All calculations employed VASP’s default three-dimensional periodic boundary conditions without vacuum layers, with the supercell replicated infinitely along x, y, and z. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary Information Reporting Summary Transparent Peer Review file Source data Source data Acknowledgements The research is supported by the ‘111 Center’ (B25033) [X.J.Z.] and the National Natural Science Foundation of China (Nos. W2511065 [X.J.Z.], 52002125 [W.H.Z.], 51802089 [H.L.L.], 62405370 [T.Z.F.]). Author contributions Junyu He carried out part of the optical experiment, data analysis, and the writing of the manuscript. Feng Jiang initiated the material growth approach at the early research stage. Jun Luo and Mengqiu Cai carried out theoretical calculations. Biyuan Zheng and Honglai Li optimized and vigorously promoted the material growth. Tongqing Sun, Chenglin He, Mengjian Zhu, and Jiahao Liu performed the optical characterizations. Shulin Chen, Yaonan Xiong, and Zhou Li performed the microstructure characterization. Tingzhao Fu, Jing Wu, completed the preparation of experimental reagents. Fang Wang assisted with the data analysis and discussion. Weihao Zheng, Zhihong Zhu and Xiujuan Zhuang proposed research ideas, participated in the experiment deployment and data analysis, completed the manuscript revisions, and provided funding support. All authors discussed the results and commented on the manuscript. Peer review Peer review information Nature Communications thanks Wenhao Shao, Guichuan Xing and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available. Data availability The source data generated in this study have been deposited in the Zenodo database: https://sandbox.zenodo.org/records/391164. The data supporting the findings of this study are also available from the corresponding author upon request. Source data are provided with this paper. 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. These authors contributed equally: Junyu He, Jun Luo, Biyuan Zheng. Contributor Information Zhihong Zhu, Email: ZZHWCX@163.com. Weihao Zheng, Email: zhengweihao@nudt.edu.cn. Xiujuan Zhuang, Email: zhuangxj@hnu.edu.cn. Supplementary information The online version contains supplementary material available at 10.1038/s41467-025-66091-2. References Associated Data Supplementary Materials Supplementary Information Reporting Summary Transparent Peer Review file Source data Data Availability Statement The source data generated in this study have been deposited in the Zenodo database: https://sandbox.zenodo.org/records/391164. The data supporting the findings of this study are also available from the corresponding author upon request. Source data are provided with this paper.