Iron deficiency strategy toward single-phase high-entropy M-type ferrites for broadband millimeter-wave absorption with low thickness Authors: Chuyang Liu, Yan Feng, Xueyu Xiang, Yujing Zhang, Er Liu, Yanyan Shao, Pan Ying, Dong-Hyun Kim, Guangbin Ji Published: 2026 ABSTRACT AbstractElectromagnetic pollution and radiation now extend to encompass millimeter waves, while the development of broadband millimeter-wave absorbers remains significantly hindered by lacking efficient loss mechanisms. High-entropy M-type ferrites are promising due to their excellent magnetic resonance and dielectric polarization, yet precise electromagnetic parameter regulation remains challenging because intermediate-phase impurities persist from insufficient configurational entropy. Herein, we propose an iron-deficient strategy to synthesize the single-phase Zr4+-doped high-entropy M-type ferrites. It is found that iron deficiency eliminates impurities by enhancing entropy and provides a more adaptable local structure. The synergistic effect of high entropy and iron deficiency amplifies the dielectric loss capacity, while Zr4+ doping broadens the magnetic loss range by modulating natural resonance. This yields an ultra-broad absorption bandwidth of 12.18 GHz and ultra-thin matching thickness below 1 mm within the millimeter-wave atmospheric window around 35 GHz, offering insights for absorbers against electromagnetic pollution and radar stealth.Subject terms: Magnetic properties and materials, Electronic properties and materials Introduction The utilization of electromagnetic waves has now been broadened to encompass millimeter waves1–4. Notably, the 35 GHz band located within an atmospheric transmission window has been widely adopted in 5 G base stations, user equipment and other wireless infrastructure owing to the diszards of low signal attenuation and strong anti-interference ability5,6. Consequently, to mitigate the growing environmental and health concerns associated with millimeter-wave radiation, it is imperative to undertake targeted research and development efforts aimed at developing highly efficient millimeter-wave absorbing materials, particularly within the 35 GHz atmospheric transmission window. As a remarkably versatile material, M-type ferrite has been widely recognized for its significant potential in the field of millimeter-wave absorption7. The high magnetocrystalline anisotropy field (Ha) engenders an intrinsic natural magnetic resonance peak within the millimeter-wave spectrum around 40 GHz. Furthermore, the resonance peak frequency can be can be precisely tuned via ion doping, thereby enabling programmable and broadband magnetic loss enhancement8,9. Nevertheless, the inherent dielectric loss within the millimeter-wave range is conspicuously insufficient, primarily attributable to its expansive bandgap characteristics and scarce polarization mechanisms10. We previously proposed an innovative high-valence ion doping methodology aimed at facilitating electron hopping between Fe3+ and Fe2+ ions in the M type ferrite, which augmented the dielectric loss to a notable extent5,11. However, the limited solubility of Fe2+ ions in the host lattice imposes a well-defined upper bound on dielectric loss enhancement achievable via this approach alone, thereby constraining the overall advancement of millimeter-wave absorption performance. Numerous researches have elucidated that high-entropy configuration can substantially reduce the bandgap of electronic band structure and induce abundant polarization behaviors in oxide compounds, thereby enhancing the overall dielectric loss12,13. Moreover, the cocktail effect inherent enables the fine-tuning of the magnetic resonance behavior through multi-component regulation, which significantly facilitates the attainment of critical objectives in optimizing impedance matching and enhancing attenuation characteristics14. Besides, the previous work has confirmed the immense scientific exploration potential of high-valence ion doping in millimeter-wave absorption5,11. Especially, the strategic selection of Zr4+ ions with lower electronegativity as dopants could promote greater ionicity in the chemical bonding to improve lattice stability. Besides, Zr4+ ions preferentially occupy 2b and 4f1 sites of Fe3+ ions associated with high magnetic anisotropy, enabling more effective modulation of the natural resonance frequency8. However, it still remains a tantalizing subject in the field of materials science for the synthesis of single-phase high-entropy M-type ferrites with selectively Zr4+ ions doping owing to their exceptionally complex crystal structure. Additionally, the commonly co-occurring impurity phase α-Fe2O3 is antiferromagnetic and exhibits low electrical conductivity, resulting in negligible contributions to both dielectric and magnetic losses and thus diminishing the millimeter-wave absorption performance. In fact, a large amount of research experience underscores that an iron-deficient formula is conducive to inhibiting the formation of common impurity phase α-Fe2O3 to generate single-phase M-type ferrites15,16. The introduction of iron deficiency not only significantly amplifies the entropy value of M-type ferrite but also cultivates a more adaptable local environment, which facilitates the defects within M-type ferrite to attain a self-consistent equilibrium and thus contributes to the mitigation of localized lattice distortion17. Herein, this study proposes an iron-deficient strategy in Zr4+-doped high-entropy M-type ferrite, with the objective of achieving an expansive absorption bandwidth and a thin matching thickness concurrently around the millimeter-wave atmospheric window of 35 GHz. Besides, our synthesis strategy is specifically designed for practical feasibility by using a conventional sol-gel method, followed by a straightforward heat treatment in ambient air. No specialized atmospheres, exotic precursors or multi-step processing are required, which demonstrates that the proposed high-entropy M-type ferrites can be synthesized through a scalable and industrially compatible route. Furthermore, first-principle calculations are conducted to elucidate the electronic band structures and differential charge density maps of the as-prepared ferrites. Micromagnetic simulations are further explored to study the influences of high-entropy and iron-deficient strategies on the magnetic characteristics. These approaches play a crucial role in elucidating the underlying mechanisms responsible for the optimized millimeter-wave absorbing performance. Ultimately, an extensive absorption bandwidth of 12.18 GHz is achieved within the atmospheric window around 35 GHz under a matching thickness less than 1 mm. These findings offer fundamental insights into the development of advanced millimeter-wave absorbers capable of simultaneously achieving broad absorption bandwidth and submillimeter structural thickness. Results and Discussion Formation of single-phase high-entropy M-type ferrite by iron deficiency strategy Figure 1a presents a detailed schematic diagram that illustrates the formation of a series of high-entropy M-type ferrites through the sol-gel method. Specifically, the crystal structures of M-type barium ferrite (BaFe12O19) and iron-deficient high-entropy M-type ferrite ((LaBaCaSrPb)0.2Fe10.5O19) are displayed in Fig. 1b. Whilst the crystal structures of Zr4+ ions doped specimens (LaBaCaSrPb)0.2Fe10.5ZrxO19 with x = 0.1 and 0.2, are illustrated in Supplementary Figs. 1–2, respectively. Gibbs free energy, enthalpy and entropy of the various M-type ferrites are calculated based on the first-principles calculation. As shown in Fig. 1c, all the calculated values of the M-type barium, lanthanum, calcium, strontium, lead ferrites remain nearly constant. However, a marked decrease in Gibbs free energy becomes evident for the high-entropy ferrite, which can be predominantly attributed to the substantially reduced formation enthalpy as well as the concomitant rise in entropy. Moreover, the iron-deficient high-entropy ferrite exhibits a further reduced Gibbs free energy, as the iron vacancies enhance structural disorder and consequently increase the configurational entropy. The implication here is that the iron deficiency strategy can favor the formation of M-type high-entropy ferrite phase12,14,18. Synthesis and structural characterizations of the various M-type ferrites. a Preparation flowchart and b crystal structures of the iron-deficient high-entropy M-type ferrites. c Gibbs free energy, enthalpy, and entropy of the various M-type ferrites. d XRD patterns of the iron-deficient high-entropy M-type ferrites measured with Cu Kα radiation. e full XPS spectra of the iron-deficient high-entropy M-type ferrites. f Synchrotron radiation XAS, XMCD and g Mössbauer spectra of the iron-deficient high-entropy M-type ferrites with x = 0 and x = 0.2. Source data are provided as a Source Data file. The X-ray diffraction (XRD) patterns (Supplementary Fig. 3) unveil distinct diffraction peaks corresponding to the α-Fe2O3 impurity phase (JCPDS NO.33-0664) in addition to those indexed to M-type ferrite phase, for the high-entropy samples synthesized with a nominal stoichiometric ratio of (Fe+Zr):(Ba+Ca+La+Sr+Pb)=12:1. If the sintering temperature is further increased in an attempt to suppress the impurity phase, the samples tend to adhere to the crucible and fail to form discrete powder particles, as observed in Supplementary Fig. 4. By contrast, XRD patterns of the high-entropy samples synthesized through an iron-deficiency strategy under identical processing conditions (Fig. 1d) are all indexed to the characteristic crystal planes of the M-type ferrite phase (JCPDS No:78-0133) without detectable impurity phase. This compelling evidence substantiates that the implementation of an iron deficiency strategy can effectively mitigate the formation of α-Fe2O3 impurity phase, thereby facilitating the attainment of single-phase high-entropy M-type ferrite. By inspection of a magnified view, the prominent diffraction peaks gradually shift towards lower angles with an increasing proportion of Zr4+ ions, which is possibly owing to larger ionic radii of Zr4+ ions (0.72 Å) occupying the sites of Fe3+ ions (0.645 Å)19,20. To evaluate the thermal stability of the synthesized high-entropy ferrite, the DSC curve of the sample with Zr4+ doping ratio of x = 0.1 is measured and compared with that of a commercial M-type barium ferrite. The results illustrated in Supplementary Fig. 5 present that the M-type barium ferrite and high-entropy M-type ferrite both reveal stable thermodynamic properties below 1400 °C. Moreover, the temperature (1449 °C) at which high-entropy M-type ferrite undergoes significant decomposition is higher than that of pure M-type barium ferrite (1431.1 °C), indicating the further enhancement of the robust thermodynamic stability for the high-entropy M-type ferrite. Full X-ray photoelectron spectroscopy (XPS) spectra shown in Fig. 1e reveal the fact that multi elements La, Ba, Ca, Sr, Pb, Fe, Zr, O coexists within the specimens. Furthermore, the actual stoichiometries are further identified using inductively coupled plasma optical emission spectroscopy (ICP-OES). The results are shown in Supplementary Table. 1, which demonstrate that the measured elemental compositions (Fe+Zr):(Ba+Ca+La+Sr+Pb) = 10.96:1, 11.08:1 and 11.13:1 basically agree with the designed compositions of (LaBaCaSrPb)0.2Fe10.5ZrxO19 (x = 0, 0.1, 0.2), agreeing well with the theoretical ratios of 10.5:1, 10.6:1, and 10.7:1, respectively. These minor discrepancies might be attributed to the elements volatilization during the sintering and the incomplete digestion process, but overall the actual ratios are all significantly lower than the standard stoichiometric ratio of 12:1. Given the detection of single-phase M-type ferrite as shown in Fig. 1b, it can be inferred that the various elements have been successfully incorporated into the M-type ferrite lattice with a controlled concentration of iron vacancies, resulting in formation of single-phase high-entropy iron-deficient M-type ferrites with partial Zr4+ ions substituting for Fe3+ ions. To investigate the valance state of Fe ions, X-ray absorption spectroscopy (XAS) of the high-entropy M-type ferrite without and with Zr4+ doping content of x = 0.2 were measured at Fe L-edge (Fig. 1f) by a total electron yield mode with 80% circularly polarized X-ray, where the L3 and L2 edge features are deconvoluted into contributions from Fe2+ (708.0 and 721.2 eV) and Fe3+ peaks (709.9 and 723.1 eV). In addition, the broad peak around 712 eV is attributed to the satellite peak of Fe8,21. The results imply that the signal of Fe2+ is pronounced while the Fe3+ contribution still maintains to be dominant (~80%), which is evidenced in both cases of L3 and L2 absorption edges. While the ratios of Fe3+ and Fe2+ appear to be scarcely affected by the doping behavior of Zr4+. In fact, the emergence of Fe2+ is mainly attributed to the unavoidable creation of positively charged oxygen vacancies during the process of high-temperature sintering22. As identified in Supplementary Fig. 6, alongside the prominent peak of lattice oxygen observed at approximately 529.7 eV, a notable peak around 531.2 eV attributable to oxygen vacancies is also distinctly detected for the XPS O1s spectra of the iron-deficient high-entropy M-type ferrites23. The oxygen vacancy content, derived from the corresponding peak areas, exhibits an obvious decreasing trend with the increase of Zr4+ doping ratio (Supplementary Table. 2), potentially facilitating the reduction of Fe2+ concentration. While from another perspective, the incorporation of positively charged Zr4+ ions into the iron-deficient high-entropy ferrite can also induce the partial transformation from Fe3+ to Fe2+ for maintaining the charge balance. Consequently, due to the synergistic effect of these two competing mechanisms, the ratio of Fe3+ to Fe2+ in the ferrite remains nearly unchanged. Moreover, X-ray magnetic circular dichroism (XMCD) has also been measured with magnetic field of +4000/−4000 Oe along the direction of X-ray, as shown in Fig. 1f. Unfortunately, it is practically difficult to apply the sum rule to separate the spin and orbital moment contribution due to the low XMCD signal at L2 edge24. However, it is noteworthy that the sample doped with Zr4+ exhibits a stronger XMCD signal from Fe3+ at the L3 edge, indicating enhanced crystallinity due to Zr4+ doping. This speculation is consistent with the XRD observation (Fig. 1d) that the diffraction peaks of the M-type ferrite phase intensify with increasing Zr4+ doping concentration, which can be explained by the Fe composition deficiency in the present sample being compensated by the additional incorporation of Zr4+ ions. From oscillatory XMCD signal at L3 edge, it is implied that Fe3+ octagonal coordination might be more abundant in the sample with Zr4+ doping25. To directly investigate the specific effect of Zr4+ ions doping on Fe3+ ions occupation distribution, Mössbauer spectra are further performed and well-fitted as displayed in Fig. 1g. The Mössbauer parameters presented in Supplementary Tables. 3–4 demonstrate that the samples both consist five magnetic sextets (12k, 4f1, 4f2, 2a, 2b) and the distribution proportions of these positions are identified to be close to the theoretical ratio of 50:17:17:8:85. While it is found that Fe3+ ions tend to migrate from trigonal site (2b) and tetrahedral site (4f1) which collectively contribute most significantly to magnetic anisotropy, toward octahedral site (4f2) upon Zr4+ doping. This can be attributed to the fact that the introduced Zr4+ ions with smaller electronegativity exhibit a preference for occupying the low-symmetry 2b and 4f1 sites5,8. Additionally, the values of isomer shifts (I. S.), quadrupole splitting (Q. S.), hyperfine field (H) and line width (Γ) experience minor fluctuations after incorporation of Zr4+ ions. The results further confirm that Zr4+ ions have successfully entered into the lattice of M-type ferrite and lead to a fine rearrangement of the iron positions distribution to occupy more octahedral positions. The microstructures of the synthesized (LaBaCaSrPb)0.2Fe10.5ZrxO19 (x = 0–0.2) are characterized by scanning electron microscope (SEM), shown in Supplementary Fig. 7. According to the observation results, the flaky particles of HER-BaM series samples exhibit a closely packed block crystal morphology. This phenomenon can be attributed to the migration of substances under high-temperature sintering conditions, which leads to the formation of robust neck connections between adjacent particles26. However, a clear trend of grain size increasing from approximately 1–10 μm to 5–20 μm with rising Zr4+ doping content from x = 0 to x = 0.1 is still evident. Moreover, characterizations by spherical aberration corrected transmission electron microscope (AC-TEM) are applied to reveal the structure details of the iron-deficient high-entropy M-type ferrites. The fast Fourier transform (FFT) pattern of the bright field TEM image (Fig. 2a) verifies the hexagonal structure of the ferrite with the zone axis of [001], as illustrated in Fig. 2b. The high-resolution TEM image presented in Fig. 2c is found to be in good agreement with the crystal structure of M-type ferrite viewed along the [001] zone axis (Fig. 2d), further confirming that iron deficiency and high-entropy strategies do not change the fundamental crystal structure of the ferrite. The high-angle annular darkfield (HAADF) imaging and energy dispersive spectroscopy (EDS) mapping (Fig. 2e) reveal a uniform distribution of La, Ba, Ca, Sr, Pb, Fe, Zr elements with Fe remaining the dominant constituent, which is fully consistent with the intended stoichiometric design. In addition, Fig. 2f, g display the TEM, high-resolution TEM images of the iron-deficient high-entropy M-type ferrite without Zr4+ doping ((LaBaCaSrPb)0.2Fe10.5O19). The geometric phase analysis (GPA) conducted along the εxx (Fig. 2h) and εyy (Fig. 2i) directions demonstrate the certain crystalline distortions induced by the combined effects of iron deficiency and high-entropy design. Such lattice distortions are found to be commonly distributed within the deformed matrixes12,27. By contrast, the TEM, high-resolution TEM images and the derived GPA mappings along εxx and εyy directions of the Zr4+ doped iron-deficient high-entropy M-type ferrite (x = 0.2) are presented in Fig. 2j–n, respectively. As observed, the doped region demonstrates markedly increased lattice strains and extensive lattice distortions resulting from Zr4+ ion doping. This structural phenomenon is likely to significantly enhance the polarization behavior of the ferrite material, thereby contributing to improved microwave absorption performance. AC-TEM characterizations of the iron-deficient high-entropy M-type ferrites. a Spherical aberration corrected bright field TEM image, b corresponding FFT diffraction patterns, c high-resolution TEM image, d crystal structure from view direction, and e HAADF image and the corresponding EDS mapping of the (LaBaCaSrPb)0.2Fe10.5ZrxO19 with x = 0.1. f Spherical aberration corrected bright field TEM image, g high-resolution TEM image of the region enclosed by the white dashed box from f, corresponding geometric phase analysis (GPA) strain mapping along the h εxx and i εyy directions of the (LaBaCaSrPb)0.2Fe10.5O19. j Spherical aberration corrected bright field TEM image, k high-resolution TEM image of the region enclosed by the white dashed box from j, corresponding geometric phase analysis (GPA) strain mapping along the l εxx and m εyy directions of the (LaBaCaSrPb)0.2Fe10.5ZrxO19 with x = 0.2. The above TEM characterizations were repeated three times independently with similar results. Modulation of multiple dielectric polarization and magnetic resonance effects through Zr4+ doping To conduct an in-depth analysis of the influence of the high-entropy and Zr4+ doping effects on microwave absorption performance, the electromagnetic parameters of HER series ferrites are measured within the 26.5–40 GHz frequency range using a vector network analyzer. As shown in Fig. 3a, it is found that the real part and imagery part of complex permittivity locate at the range of 4.5–7.5, 1.8–4.5 respectively, which are obviously higher than those of the previous Ti4+, Zr4+, Nb5+ single doped and La3+-Zr4+ co-doped M-Type barium ferrites (Supplementary Fig. 8)5,8,28,29. As far as we can discern, conduction loss exhibits a profound and intrinsic correlation with electrical conductivity, which can be qualitatively evaluated through the analysis of the electronic band structure. Based on the First-principles calculation results (Fig. 3d), M-type barium ferrite presents a typical semiconductor structure with band gap of 1.576 eV, and the gap can be reduced to 0.645 eV by iron-deficient strategy. Similar results are also obtained in the other M-type ferrites (Supplementary Figs. 9–12), interpreting the positive role of iron-deficient strategy in elevating the electrical conductivity and thus the complex permittivity. More importantly, further high-entropy design prompts a significant transformation of the electronic structure towards conductive properties. The results of room-temperature DC resistivity measurements shown in Supplementary Fig. 13 reveal that the iron-deficient high-entropy Zr4+-doped M-type ferrite powders exhibit decreased resistivity values by 3 to 4 orders of magnitude compared with M-type barium ferrite, which is consistent well with the DFT calculation results and identified as a key contributing mechanism to the dramatic increase of complex permittivity in the M-type ferrites. On the other hand, once polarization relaxation occurs, it will simultaneously affect the values of ε′ and ε″ in the specific frequency range30. Based on Debye theory expressed as Eq. 1, relaxation processes can be elucidated by ε‘-ε“ Cole-Cole curves, where one single circle indicates one type of Debye relaxing process. The ε∞ and εs represent the relative permittivity at infinite frequency and static permittivity, respectively. According to Fig. 3b, two distorted semicircles appear in the ε′-ε″ Cole-Cole curves for all the samples, indicating the existence of two main Debye relaxation processes during the 26.5–40 GHz within the multi-component system of high-entropy materials31,32. Overall, the abundant dipole polarization relaxation and electronic band structure rearrangement caused by the high-entropy effect collectively enhance the contributions of polarization and conduction, thereby increasing the complex permittivity. Moreover, to investigate how Zr4+ ions doping affects the dielectric loss mechanism, the polarization loss and conductance loss of the high-entropy series samples are quantitatively analyzed by nonlinear least squares fitting calculation. The specific calculation processes have been elaborately introduced in our previous research work31. As can be seen in Fig. 3c, the conductance loss slightly decreases with the increase of Zr4+ doping ratio. This result is most likely due to the fact that the concentration of oxygen vacancies decreases with the amount of Zr4+ doping increasing as shown in Supplementary Fig. 6, thereby leading to a slight reduction in the electrical conductivity of the samples, which in turn mitigates conductance loss33. Besides, polarization loss also exhibits an obvious decreasing trend by increasing Zr4+ doping in the range of 26.5–33 GHz, which is likely attributed to the incorporation of Zr4+ cations compensating for dipoles and defects induced by Fe vacancies. However, it is intriguing to find that polarization loss shows an obvious upward trend by Zr4+ doping within the high frequency range of 37–40 GHz. To visualize the point, differential charge density maps of a series of M-type ferrites are calculated based on the density functional theory. Initially, differential charge density maps of M-type barium, calcium, lanthanum, strontium, lead ferrite and their iron-deficient phases are displayed in Supplementary Fig. 14. The results evidently illustrate that Fe vacancy could induce severe electron cloud distortions in the surrounding ions across all kinds of M-type ferrite34,35. Moreover, intrinsically severer electronic cloud distortions are observed in the high-entropy M-type ferrite (Fig. 3e) in comparison with the low-entropy M-type ferrites (Supplementary Fig. 14). The findings compellingly reveal the presence of a wealth of polar dipoles within the high-entropy systems. Similarly, Fe vacancy further dramatically aggravates the electron cloud distortions of the adjacent ions in the high-entropy M-type ferrite, promoting charge separation to create dipole moments. Moreover, the introduction of Zr4+ is observed to effectively disrupt the local electron distribution, leading to the formation of new dipoles and thus significantly enhancing polarization responses and the associated relaxation processes. This provides a powerful evidence for the increased polarization loss (Fig. 3c) at high frequency range of 37–40 GHz for Zr4+-doped samples. Dielectric properties of the iron-deficient high-entropy M-type ferrites. a The complex permittivity of the (LaBaCaSrPb)0.2Fe10.5ZrxO19 with x = 0–0.2. b The ε′-ε″ Cole-Cole curves, c conductivity loss and polarization loss obtained by non-linear least squares fitting method of the (LaBaCaSrPb)0.2Fe10.5ZrxO19 with x = 0–0.2. d Electronic band structure of the M-type barium ferrite, iron-deficient barium ferrite and iron-deficient high-entropy ferrite. e Differential charge density maps of the high-entropy, iron-deficient high-entropy, Zr4+-doped iron-deficient high-entropy M-type ferrite calculated based on the density functional theory. Source data are provided as a Source Data file. Meanwhile, as shown in Fig. 4a, b, both the real and imaginary parts of complex permeability exhibit asymmetric resonance peak characteristics, and these peaks gradually shift to lower frequency ranges as the doping ratio of Zr4+ increases. In fact, the dominant magnetic loss of M-type ferrite at dozens of GHz is attributed to the natural resonance loss, whose frequency (fr) is directly proportional to the magnetocrystalline anisotropy field (Ha) as displayed in Eq. 25,8: where γ is gyromagnetic ratio and g is Lande´ factor. Based on the reported literatures, the magnetocrystalline anisotropy of M-type ferrites primarily originates from the magnetic Fe3+ ions located within the oxygen polyhedra36,37. As evidenced by the analysis of Mössbauer spectra, the incorporation of non-magnetic Zr4+ ions induces a preferential redistribution of Fe3+ ions from trigonal and tetrahedral sites contributing higher Ha to octahedral sites exhibiting lower Ha, which reduces the overall effective anisotropy field and consequently lowers the natural resonance frequency as described by Eq. 15,8. Additionally, the intensity of the resonance loss peak is observed to decrease slightly with increasing Zr4+ content in Fig. 4b, which can likely be attributed to the reduced saturation magnetization (Ms) shown in Supplementary Fig. 15a. This is due to the fact that Zr4+ ions dominantly occupy the Fe3+ ions at the tetrahedral 4f1 position with up-spin38. Since the saturated magnetization of M-type ferrites primarily arises from the net magnetic moment of Fe3+ ions at five crystallographic sites, the various cations at the M-sites have minimal impact on saturation magnetization39. The saturation magnetization of the designed (LaBaCaSrPb)0.2Fe10.5O19 is also compared with that of single-phase BaFe12O19 sintered under identical conditions. As shown in Supplementary Fig. 15a, a significant reduction of approximately 18.6% is observed, decreasing from ~59 emu/g for BaFe12O19 to ~48 emu/g for (LaBaCaSrPb)0.2Fe10.5O19. This marked decrease in saturation magnetization should be primarily attributed to the reduced content of magnetic Fe3+ ions, further confirming the formation of iron vacancies in the designed high-entropy ferrites. Moreover, the dM/dH–H plots (Supplementary Fig. 15b) all reveal only a single peak with wide range and high intensity. This observation supports the successful preparation of single-phase high-entropy M-type ferrites40. Besides, the contribution by eddy current loss can be analyzed through the relationship of f and μ“(μ‘)-2f-1. As shown in Fig. 4c, the samples with different proportions all present curves with distinct single peaks in the frequency range of 26.5–40 GHz, rather than straight lines that do not vary with the frequency band. These results indicate that eddy current losses are negligible, and natural resonance is the dominant mechanism contributing to magnetic loss in the high-entropy M-type ferrite under the present experimental conditions41. Magnetic properties of the iron-deficient high-entropy M-type ferrites. a μ‘, b μ“ and c μ“(μ‘)-2f-1 of the (LaBaCaSrPb)0.2Fe10.5ZrxO19 with x = 0–0.2. The μ‘-μ“ Cole-Cole curves of the (LaBaCaSrPb)0.2Fe10.5ZrxO19 with d x = 0, e x = 0.1 and f x = 0.2. g Peak separation fitting for μ“-f spectra of the (LaBaCaSrPb)0.2Fe10.5ZrxO19 with x = 0–0.2. h Magnetic force microscopy images (2D) of the (LaBaCaSrPb)0.2Fe10.5ZrxO19 with x = 0 and x = 0.2. i Micromagnetic simulations of the high-entropy iron-deficient M-type ferrite with and without Zr4+ doping. Source data are provided as a Source Data file. The μ‘-μ“ Cole-Cole curves are generally depicted to analyze the magnetic resonance behavior according to the following Eq. 3: Where Bm and Hm represent the maximum magnetic induction and maximum magnetic field strength, respectively. On the basis of Xing’s theory, a semicircle in the μ‘-μ“ plots corresponds to one distinct type of magnetic resonance mechanism42. To the best of our knowledge, natural resonance represents the primary magnetic resonance mechanism of M-type ferrite in the frequency range of several dozen GHz5,8. The presence of multiple semicircular arcs in Fig. 4d–f implies that all high-entropy M-type ferrites exhibit multiple natural resonances within the frequency range of 26.5–40 GHz. For further investigating the intriguing phenomenon, the asymmetric resonance peak in the μ“-f plots are resolved and presented in Fig. 4g. For the iron-deficient, high-entropy M-type ferrite without Zr4+ doping, the asymmetric resonance peak can be deconvoluted into three distinct resonance regions located at <26.5–30.5 GHz (region Ⅰ), 30.5–38.0 GHz (region Ⅱ) and 38.0 ~ > 40.0 GHz (region Ⅲ) respectively, which are likely attributable to intrinsic natural resonances arising from the compositionally heterogeneous high-entropy components43. Furthermore, the resonance peak in region Ⅱ can be further decomposed into two distinct peaks, which are contributed by new g-factor generated through exchange coupling between Fe3+ and Fe2+ ions (>2.0) and the intrinsic g-factor of Fe³⁺ ions (2.0) separately, as described by Eq. 128. Additionally, the frequencies of the three resonance regions are observed to gradually decrease with increasing Zr4+ content. As can be seen, new resonance peaks emerge in region Ⅲ with typically Fe3+-Fe2+ exchange coupling characteristics at 33.7 ~ > 40.0 GHz when the Zr4+ content reaches x = 0.2, a phenomenon that may be ascribed to a frequency shift originating from above 40 GHz. Above all, under the synergistic effects of high-entropy compositional design and Fe3+-Fe2+ exchange coupling effect, various natural resonance peaks of high-entropy M-type ferrites coalesce and gradually transform into a continuous magnetic loss profile, resulting in significantly broadened magnetic loss range. The Zr4+ doping can further effectively modulate both the number and frequency of natural resonance peaks in the high-entropy ferrites, thereby enabling precise regulation of the magnetic loss capacity within the 26.5–40 GHz frequency range. The magnetic domain structures of the high-entropy M-type ferrites are further detected by magnetic force microscopy (MFM). The sample preparation follows the standard MFM procedures for magnetic powders: ultrasonic dispersion in isopropanol, deposition on a clean silicon wafer, natural drying in a contamination-free environment, and removal of loose particles using a dust remover. As illustrated in Fig. 4h and Supplementary Fig. 16, the red area indicates that the magnetization direction of the sample aligns with that of the probe, whereas the blue area presents antiparallel orientation. By comparison, the red/blue regions of the high-entropy sample without Zr4+ doping are smaller and more dispersed, while the regions become larger and more concentrated for the sample with Zr4+ doping ratio of x = 0.2. This observation indicates that Zr4+ doping leads to an increase in the average magnetic domain size, which is probably attributed to the substitution of non-magnetic Zr4+ ions for Fe3+ ions by reducing Ha of the M-type ferrite. In addition, micromagnetic simulations are performed by Mumax3 to visualize the spatial distribution of magnetic moments across the cross section under a 30 GHz electromagnetic excitation, and the color maps encode the in-plane orientation of magnetic moments. Regarding to the pristine M-type barium ferrite, the shift, vanishment and reappearance of magnetic vortices subsequently occur under the influence of the external electromagnetic field as shown in Supplementary Fig. 17. In comparison, the high-entropy M-type components result in more pronounced magnetic vortices, as evidenced by the distinct color differences in magnetic moment directions (Supplementary Fig. 18). Moreover, it is clearly observed in Fig. 4i that the iron-deficient strategy enables the magnetic vortex of high-entropy M-type ferrite to reform over the same time period, and the additional doping with Zr4+ ions induces a significant displacement of the magnetic vortex. Theoretically, substantial local torques, arising from strong exchange fields in the magnetic vortex core region during dynamic motion or collapse, drive rapid energy dissipation into the surrounding magnetic medium via the excitation and propagation of spin waves44. Therefore, the more pronounced and rapidly reconfigurable magnetic vortex structures can broaden the effective magnetic response spectrum, matching the enlarged magnetic loss peak in the experimental μ“ spectra (Fig. 4b). Consequently, the overall microwave absorption performance could be obviously improved by the reinforced magnetic loss capability enabling effective decoupling of impedance matching and microwave dissipation45. While it should be noted that the micromagnetic simulations are performed at 0 K under ground-state conditions. Nevertheless, thermal fluctuations at room temperature do not substantially alter the stability of the magnetic vortex state or the fundamental mechanisms governing its field-driven dynamics. This robustness stems from the high Curie temperature of M-type hexaferrite (~730 K), which is more than seven times the ambient temperature for ensuring negligible thermal perturbation to its magnetic ordering under operational conditions. Thus, the results can still serve as an important reference for analyzing the magnetic behavior of M-type ferrite. Mechanisms and applications for the efficient millimeter-wave absorption capacity The reflection loss (RL) value is a standard metric for evaluating the microwave absorption performance and is calculated based on the measured electromagnetic parameters within a specific thickness range, using the metal backplane model as described in Eqs. 2–346,47. In general, a lower RL value indicates stronger absorption of electromagnetic waves. The 3D RL curves for the high-entropy series samples within the frequency range of 26.5–40 GHz and thickness range of 0.6–5 mm are calculated and presented in Fig. 5a. At the bottom projection, the absorption capacity of the material for electromagnetic waves under different thicknesses can be clearly seen. The black line area represents the range of RL < −10 dB, which indicates that the absorption rate for electromagnetic waves exceeds 90%. It can be found that the samples with doping ratios of x = 0.1 and x = 0.2 exhibit effective absorption bandwidths that nearly cover the entire frequency range under a calculated low thickness of approximately 1 mm, particularly within the thickness range of 0.80–0.95 mm. For an intuitive comparison, the 2D RL curves under calculated thicknesses of 0.80 mm, 0.85 mm, 0.95 mm, and the corresponding impedance matching curves as well as effective absorption bandwidth (EAB) column charts are presented in Fig. 5b. Specifically, the high-entropy ferrite without Zr4+ doping has the minimum RL (RLmin) value of −41.99 dB at a thickness of 0.8 mm. In addition, the EAB reaches the maximum value of 7.83 GHz at a thickness of 0.95 mm. When the Zr4+ doping ratio is x = 0.1, the RLmin further intensifies to −45.98 dB under a thickness of 0.85 mm. The sample can achieve a maximum EAB value of 11.10 GHz with a matching thickness of 0.95 mm. Further elevating the Zr4+ doping ratio to x = 0.2, an RLmin of −31.31 dB and an EAB of 12.18 GHz around the atmospheric window of 35 GHz can be simultaneously obtained at a thickness of 0.95 mm, covering 90.2% of the entire frequency band. It is evident that the Zr4+-doped iron-deficient high-entropy M-type ferrites fully satisfy the millimeter-wave absorption requirements within the ~35 GHz atmospheric window. Microwave absorption properties of the iron-deficient high-entropy M-type ferrites. a 3D RL curves, b 2D RL curves, |Zin/Z0| values under thicknesses of 0.80 mm, 0.85 mm, 0.95 mm and EAB bar charts of the (LaBaCaSrPb)0.2Fe10.5ZrxO19 with x = 0-0.2. c Schematic diagram of the millimeter-wave absorption mechanisms. This is a schematic, not actual experimental data. Source data are provided as a Source Data file. To unveil the reinforced millimeter-wave absorption performance, the input impedance matching characteristics of the as-designed samples under different thicknesses are calculated. It can be found that the |Zin/Z0| values gradually decrease with thickness increasing from 0.80 to 0.95 mm for all samples. Obvious RL peaks appear at the corresponding frequencies where the |Zin/Z0| values approach to ~1, indicating that a significant portion of the electromagnetic waves enters the material for effective absorption. Both RL and EAB of the Zr4+ doped sample are improved compared to the sample with x = 0. This is likely attributable to Zr4+ doping modifying the lattice distribution of Fe3+ ions in M-type ferrite, thereby reducing the magnetic anisotropy field and shifting the magnetic resonance toward lower frequencies, as previously discussed. Consequently, the |Zin/Z0| values approach ~1 across a broader frequency range for the enhanced absorption properties. Moreover, it should be noted that the presence of multiple magnetic loss mechanisms does not guarantee that they will manifest as visibly distinct, separate peaks in RL curves at every thickness in the corresponding frequency range. At a certain thickness, if the electromagnetic parameters satisfying |Zin/Z0| values approaching ~1 are concentrated within a single frequency region, only one distinct RL peak will form. Conversely, at other thicknesses, if the suitable electromagnetic parameters are distributed across different frequency intervals, multiple distinct RL peaks can emerge. As evidenced in Fig. 5b, the existence of multi RL peaks in the sample with x = 0.1 becomes more apparent at 0.80 mm compared to the other thicknesses where the matching condition changes. This demonstrates that thickness is also a key parameter for regulating the microwave absorption performance. The intricate millimeter-wave absorption mechanism of the iron-deficient high-entropy M-type ferrite by Zr4+ doping is exquisitely illustrated in Fig. 5c, offering a visually compelling and detailed depiction of the underlying physical processes. Essentially, the millimeter-wave absorption property of the ferrites stem from the synergistic interaction between abundant magnetic loss mechanisms arising from multiple natural resonances and sufficient dielectric loss mechanisms contributed by polarization and conduction couplings, enabling millimeter waves to penetrate deeply into the ferrites and be efficiently dissipated within them. Concretely, the coexistence of multiple elements within the system inherently gives rise to multiple natural resonance peaks, which can be preferentially shifted into the target frequency band around the 35 GHz atmospheric window through non-magnetic Zr4+ ion doping. Moreover, the exchange coupling between Fe3+ and Fe2+ ions results in new g-factors distinct from those of pure Fe3+, potentially inducing the formation of additional natural resonances. These mechanisms collectively confer superior magnetic loss capability on the ferrites over a broad frequency range, spanning up to the target millimeter-wave band. On the other hand, the high-entropy compositional design facilitates the transformation of the electronic band structure of M-type ferrite from a semiconducting to a conducting state, significantly enhancing the conduction loss. Concurrently, the high-entropy constituent, iron vacancy and Zr4+ ions doping can induce substantial dipole polarizations within the system. With Zr4+ doping, though the conduction loss exhibits a slight decrease across the entire measured frequency range, the polarization loss is significantly enhanced in the high-frequency region of 37–40 GHz. This key microwave attenuation mechanism complements the degradation of conduction and magnetic losses. As a result of the optimized polarization loss arising from Zr4+-induced charge redistribution, both substantial magnetic loss and dielectric loss are achieved as shown in Supplementary Fig. 19. This contributes to improved input impedance matching, thereby enabling excellent absorption performance over a broader frequency range in the Zr4+-doped high-entropy M-type ferrites, as observed in Fig. 5b. In the practical application of radar stealth, the electromagnetic attenuation capability of absorbing materials in the far field is usually measured by the radar cross section (RCS)48–50. In this work, a PEC plate with dimensions of 110 mm × 110 mm × 0.07 mm is constructed using CST Studio Suite 2018, and the electric field distributions of the PEC plate with and without covering the as-prepared absorbing material under thickness of 2 mm are simulated, respectively. In Fig. 6a–d, the Z-axis is the incident direction of the electromagnetic wave, and the detection angle is in the range of −90° to 90°, the excitation frequency is meticulously selected at 35 GHz. The intensity change of the scattered signal can be observed through the change of color and radiation structure. It can be found that the planes coated with the absorbing materials exhibit lower scattered signals compared to the original PEC plate, and the Zr4+-doped high-entropy ferrites demonstrate reduced scattered signals relative to the sample without Zr4+ doping, which aligns with the results of the 3D RL curves. In addition, the 2D curves of the RCS values in the range of −90° to 90° are shown in Fig. 6e. Clearly, when the electromagnetic wave is incident perpendicular to the model surface (Theta = 0°), the proportion of the reflected electromagnetic wave is relatively high. However, as the detection angle gradually shifts toward 90° and −90°, the RCS value progressively decreases and exhibits multiple fluctuations, which is less than −10 dB/m² in the range of −90° to −11° and 11° to 90° through the utilization of the high-entropy ferrites. As presented in Supplementary Figs. 20–21, similar results are also obtained at the frequencies of 32 GHz and 38 GHz, which demonstrate the stability of RCS reduction behavior of our design within a certain bandwidth. Applications of the iron-deficient high-entropy M-type ferrites. The RCS values simulated by CST of the PEC plate at 35 GHz a without covering absorbing materials, and covered by (LaBaCaSrPb)0.2Fe10.5ZrxO19 with b x = 0, c x = 0.1 and d x = 0.2. e 2D RCS curves and of the pure PEC plate, and the PEC plate covered by (LaBaCaSrPb)0.2Fe10.5ZrxO19 with x = 0 (S1), x = 0.1 (S2) and x = 0.2 (S3). f Relative RCS reduction values of (LaBaCaSrPb)0.2Fe10.5ZrxO19 with x = 0-0.2 at specific angles. g The working principle schematic diagram for electromagnetic wave transmission. h Performance comparison chart between this work and the existing state-of-the-art millimeter-wave absorbers. Source data are provided as a Source Data file. Meanwhile, the relative RCS reduction values of the PEC plates equipped with high-entropy ferrites are calculated at selected angles comparing with those of the pure PEC plate, and the results are shown in Fig. 6f. When the detection angle is 12°, the relative RCS reduction value of the plate with Zr4+ doped high-entropy ferrite at x = 0.1 can reach 31.08 dB/m², and the values are always greater than that of the sample including high-entropy ferrite without Zr4+ doping at any angle. Moreover, the plate coated with Zr4+-doped high-entropy ferrite at x = 0.2 exhibits a more significant RCS reduction at higher angles, effectively attenuating scattered and reflected millimeter waves from the coating surface. Therefore, the simulation results indicate that Zr4+-doped iron-deficient high-entropy M-type ferrites are well-suited for application in the field of RCS reduction, and the schematic diagram of working principle is depicted in Fig. 6g. To unveil the profound origins underlying this exceptional millimeter-wave absorbing capability, an intuitive comparison is performed in Fig. 6h. As can be clearly seen, the millimeter-wave absorption performance of the as-prepared ferrites in this study is evidently superior to that of existing state-of-the-art low-entropy millimeter-wave absorbers in terms of both absorption bandwidth and matching thickness5,28,51–57. Additionally, the RL values of the high-entropy M-type ferrites designed without iron-deficiency (LaBaCaSrPb)0.2Fe12ZrxO19 with x = 0-0.2 are presented in Supplementary Fig. 22. By contrast, they all exhibit inferior millimeter-wave absorption performance characterized by narrower bandwidth or greater matching thickness, which can be primarily attributed to insufficient dielectric and magnetic losses as shown in Supplementary Fig. 23, arising from the antiferromagnetic impurity phase α-Fe2O3 with low conductivity (Supplementary Fig. 3) and the unfavorable magnetic resonance range. Above all, both high-entropy configuration and iron-deficient composition represent indispensable strategies for achieving exceptional millimeter-wave absorption performances, which jointly endow them with remarkable competitiveness in the demanding real-world scenarios across a broad spectrum of practical applications. In summary, single-phase Zr4+-doped high-entropy M-type ferrites have been successfully synthesized via a facile sol-gel method by employing an iron-deficiency strategy. The high-entropy and iron-deficient effects induce diverse polarization responses and trigger a transformation of electronic band structure from semiconducting to conducting, thereby synergistically enhancing the contributions of both polarization and conduction losses. Moreover, the incorporation of Zr4+ ions into M-type ferrite induces a precise rearrangement of Fe3+ ions distribution, which effectively modulates multiple natural resonance peaks to produce an enhanced magnetic loss performance within the targeted 26.5–40 GHz frequency range. Ultimately, benefiting from the synergistic effect of significant magnetic and dielectric loss mechanisms, broad absorption bandwidth of 12.18 GHz is achieved within the millimeter-wave atmospheric window near 35 GHz, along with a notable reflection loss of −45.98 dB at a thin matching thickness less than 1 mm. These findings offer valuable insights into the development of advanced millimeter-wave absorbers, with the objective of achieving the synergistic combination of lightweight design, broad bandwidth, and strong reflection loss performance. Methods Materials Zirconium nitrate pentahydrate (Zr(NO3)4·5H2O), iron nitrate nonahydrate (Fe(NO3)3·9H2O) and barium nitrate (Ba(NO3)2) were purchased from Sinopharm Chemical Reagent Co., Ltd. Lanthanum nitrate hexahydrate (La(NO3)3·6H2O), strontium nitrate (Sr(NO3)2), Calcium nitrate tetrahydrate (Ca(NO3)2·4H2O), lead acetate trihydrate (Pb(CH3CO2)2·3H2O), citric acid (C6H8O7·H2O) and ammonia water (NH3·H2O) were bought from Aladdin (Shanghai) reagents Co., Ltd. All commercial reagents utilized in this study were of analytical grade and employed without any additional purification. Synthesis of Zr4+-doped Fe-deficient high-entropy M type ferrite The Fe-deficient high-entropy M type ferrite with Zr4+ partly substituting for Fe3+ was fabricated by a traditional sol-gel approach. Firstly, 0.8658 g La(NO3)3·6H2O, 0.5227 g Ba(NO3)2, 0.4233 g Sr(NO3)2, 0.4723 g Ca(NO3)2·4H2O, 0.7587 g Pb(CH3CO2)2·3H2O and 42.42 g Fe(NO3)3·9H2O, together with 0 g Zr(NO3)4·5H2O for the x = 0 sample, 0.4293 g Zr(NO3)4·5H2O for the x = 0.1 sample and 0.8586 g Zr(NO3)4·5H2O for the x = 0.2 sample, were weighed according to the chemical composition of (LaBaCaSrPb)0.2Fe10.5ZrxO19 (x = 0-0.2). Then, they are dissolved in 300 mL deionized water under magnetic stirring to form a clear solution. Subsequently, 43.17 g citric acid for the x = 0 sample, 43.30 g citric acid for the x = 0.1 sample and 43.42 g citric acid for the x = 0.2 sample were added to the above solutions respectively in order to facilitate an adequate complexation reaction. Afterwards, ammonia water was used to adjust the pH value to 7 and stirred at 80 °C for 3 h to obtain wet gels. They were then placed in an oven at 120 °C for a duration of three days to facilitate the evaporation of the water solvent, resulting in the transformation into dry gels. The resulting dry gels were pre-calcined on a heating platform at 400 °C for 15 min. After cooling, the products were ground in a mortar to obtain reddish-brown precursor powders. Eventually, the obtained precursors were placed in a muffle furnace and subjected to a multi-stage sintering process: they were first heated to 800 °C with a rate of 10 °C/min and held for 4 h, then ramped to 1000 °C by 5 °C/min for 4 h, and finally heated to 1400 °C by 5 °C/min and kept for 3 h to achieve the desired Fe-deficient high-entropy M-type ferrite. Material characterizations X-ray diffraction (XRD, model Bruker D8 ADVANCE) was utilized to identify the crystal structure and X-ray photoelectron spectroscopy (XPS, PHI 5000 Versa probe) was performed to analyze the element states. Fe L2,3 edge XAS/XMCD spectra were collected in total-electron-yield mode, which were carried out on SINS beamline at the Hefei Light Source. The morphologies were characterized by scanning electron microscope (SEM, Hitachi S-4800), transmission electron microscope (TEM, JEOL ARM200F). The magnetic domain structures were analyzed by magnetic force microscope (AFM, Shimadzu SPM-9700HT). Hysteresis loops were measured by a magnetic property measurement system (MPMS-XL-5). The electromagnetic parameters in the range of 26.5–40 GHz (Ka band) were carried out using a vector network analyzer (Ceyear 3672 C), wherein the test samples are made by high-entropy M type ferrite and paraffin wax with mass ratio of 8:2. These mixtures were subsequently pressed into toroidal-shaped specimens for measuring electromagnetic parameters with an inner diameter of 3.04 mm, an outer diameter of 7.00 mm, and a thickness ranging from 2.0 to 3.0 mm. Gibbs free energy, differential charge density and electronic band structure calculations The differential charge density maps and electronic band structures of the various M-type ferrites were calculated based on density functional theory (DFT) by using Materials Studio software. The DFT calculations in this work were performed using the experimentally determined room-temperature crystal structure, which aim to elucidate the intrinsic electronic-structure-based mechanisms relevant to the observed macroscopic properties. The electron-ion interaction and the exchange-correlation function were based on the Generalized-Gradient-Approximation (GGA) with the Perdew-Burke-Ernzerhof (PBE) generalized gradient approximation. On the basis of the experimental results, different types of M-type ferrite were applied as calculated models using their single cells. A Monkhorst–Pack grid size of 5 × 5 × 1 was employed for surface calculations. The cut-off energy for plane wave expansions was set at 571.40 eV. Convergence tolerance optimization was performed for energy with values set at 1.0 × 10−6 eV. Gibbs free energy values of the various M-type ferrites were calculated by a combination of DFT calculation and manual calculation through the following Eq. 6, where formation enthalpy was obtained by DFT calculation while configurational entropy was manually calculated on the basis of Eq. 7. where R, N(M), and xi refer to the gas constant, atomic species, and contents at the cation sites, respectively. For the Fe deficient systems, the vacancy is also treated as a distinct species on the cation sublattice. The optimized atomic coordinates, output structures, total energies of the corresponding computational models and other relevant computational data are detailedly provided in Supplementary Data 1. Micromagnetic simulations Micromagnetic simulations were executed by Mumax3 software, aimed to evaluate the dynamic evolution process of magnetic vortices and analyze their behaviors by construction of material spheres with a radius of 200 nm. The damping factor of spin precession was 0.01, and the external magnetic field frequency was 30 GHz. The attached software Muview was applied to visualize the results calculated by Mumax3. RCS simulations RCS simulations of the PEC plate with size of 110 mm × 110 mm × 0.07 mm were carried out using CST Studio Suite 2018 software package with and without the high-entropy M type ferrite coatings. The electric boundary was established as an open boundary condition, while the excitation frequency was meticulously selected at 32, 35 and 38 GHz. The Z-axis represents the incident direction of the electromagnetic wave, and the detection angle was taken within the range from −90° to 90°. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary information Description of Additional Supplementary File Supplementary Data 1 Reporting Summary Transparent Peer Review file Source data Source data