A water-induced self-assembly approach to 3D hierarchical magnetic MXene networks for enhanced microwave absorption

Abstract

Three-dimensional (3D) network structures assembled from two-dimensional (2D) MXene nanomaterials hold enormous potential for microwave absorption (MA) due to strong dielectric loss ability. However, the uncontrollable oversized conductive paths caused by self-stacking issues and the natural lack of magnetic loss hinder the further MA application of the assembled 3D MXene. Herein, we report a facile water-induced self-assembly method to develop a series of 3D hierarchical magnetic nanocrystal@C@MXene hybrids with tunable network structures. The design of the hierarchical structure can isolate each electromagnetic component, thereby ensuring the efficient synergy of the high-density ultrafine magnetic components and high-loss dielectric components. The resulting massive heterointerfaces greatly enhance the interface polarization and reinforce the dielectric loss capability. Moreover, through control over the MXene load, the size of 3D networks can be precisely regulated to adjust the impedance matching. According to experimental and finite element simulation, all the 3D hierarchical magnetic hybrids were found to possess impressive MA properties. Specifically, ZnFe₂O₄@C@MXene exhibits a minimum reflection loss value of −62.59 dB over an effective absorption bandwidth of 4.42 GHz at a thickness of only 1.33 mm. This work provides a flexible route for constructing 3D hierarchical network materials to realize highly efficient MA performance with thin thickness.

---

Introduction

With the rapid development of the information era and the high integration of mobile terminals, ultrathin and highly efficient microwave absorption (MA) materials that can cope with electromagnetic wave pollution are urgently needed.^1,2 In general, ideal MA materials are the ones with good impedance matching and excellent attenuation capabilities, which can be constructed by the rational design of MA microstructures.^3,4 Owing to the flexible structure and excellent dielectric loss capability, the emerging two-dimensional (2D) transition metal carbide/carbonitride materials (MXenes) have been employed as a high-loss component for the microstructural designs of advanced MA materials.^5,6 The rich functional groups and charges of the surface induced by the exfoliation process endow MXenes with an excellent assembly ability, which enables MXenes to be freely assembled into various customizable structures, such as spherical structures, layered structures, and three-dimensional (3D) network structures.^7–10 In particular, the 3D network structures comprising low-density nanosheets are attractive due to their rich conductive loss networks that can prolong the attenuating paths.^11,12 However, 3D networks are prone to form oversized conductive paths due to the natural self-restacking of MXenes, resulting in the skin effect that hinders impedance matching.^13–16 Besides, the complex assembly processes and the lack of magnetic losses further limit their MA performance.¹⁷

An effective solution strategy to enhance electromagnetic synergy is to incorporate magnetic components to build 3D network structures through interfacial engineering.^18,19 Apart from introducing magnetic losses, the interfacial engineering strategy can anchor MXenes on the heterogeneous interface and construct a controllable conductive network.²⁰ The typical method for forming MXene-based network structures is to assemble MXenes with other components via hydrogen bonds or electrostatic force.²¹ However, normal magnetic materials cannot provide an active surface for assembly.²² Accordingly, some strategies based on surface modification have been exploited. These strategies rely on organic ligands, like cetrimonium bromide (CTAB), oleylamine (OA), and poly(diallyldimethylammonium chloride) (PDDA), to assist the self-assembly of magnetic structures and MXene.^23–26 Despite the fact that various magnetic components were successfully assembled with MXene through these time-consuming strategies, the resulting irregular network structure is unfavorable for MA performance.²⁷ Therefore, it remains a significant challenge to fabricate a controllable 3D magnetic network through a more feasible and simpler approach.

Proper selection of magnetic components will aid in the construction of 3D hierarchical structures and the enhancement of MA performance.¹⁹ The magnetic nanoparticles are the preferred choice due to their simple synthesis and high magnetic loss.²⁸ Meanwhile, reducing the size of magnetic nanoparticles, which increases the active surface area for assembling with MXenes, can also enhance interface loss and reduce the local accumulation of MXenes.⁹ Nevertheless, ultrafine magnetic nanoparticles of small size tend to assemble on relatively larger-sized nanosheets instead of isolating MXenes and forming networks.^29,30 Besides, the self-aggregation of ultrafine nanoparticles further exacerbates the self-restacking of MXenes, weakening the advantages of magnetic nanostructures and 3D networks.³¹ Accordingly, the construction of controllable MXene networks with high-density and uniformly dispersed magnetic nanocomponents still needs more investigations in depth.

Herein, we provide a general synthesis method of 3D hierarchical magnetic nanocrystal@C@MXene hybrids, including isolation structures composed of ultrafine magnetic nanocrystals densely isolated in spherical carbon matrixes and the controllable MXene networks anchored on matrix interfaces. The orderly isolation can efficiently coordinate the superiority of each electromagnetic component, while massive heterointerfaces created by the continuous isolation structure can greatly enhance the interface polarization and reinforce the dielectric loss capability. More importantly, the involved water-induced self-assembly method, which not only utilizes water as a solvent but also activates and participates in the assembly, can assemble 2D nanosheets into 3D MXene networks. Therefore, the size of networks can be orderly regulated, thus enabling effective modulation of impedance matching and attenuation paths. As expected, the hierarchical hybrids with tunable MXene networks were found to possess impressive MA properties. Specifically, all obtained 3D hierarchical ZnFe₂O₄@C@MXene materials with different sized networks show excellent MA performance. Moreover, regulating the size of network structures can precisely adjust the impedance matching and prolong the attenuation paths, thus greatly promoting the MA performance at ultrathin thicknesses. ZnFe₂O₄@C@MXene with 3D conductive networks delivers a minimum reflection loss (RL_min) value of −62.59 dB at 14.16 GHz over an effective absorption bandwidth (EAB, RL < −10 dB) of 4.42 GHz at an ultrathin thickness of only 1.33 mm. In short, the present work provides a flexible route for the structural design of 3D hierarchical network materials and promotes the further development of ultrathin MA materials.

Results and discussion

The general strategy for the formation of hierarchical magnetic nanocrystal@C@MXene is shown schematically in Fig. 1. The vital step for the synthesis of hierarchical structures is the water-induced self-assembly of MXene and metal–glycerate (denoted as M–G, M: magnetic metallic elements and bimetallic elements). In deionized water without adding any ligands, the surface of dispersed M–G is spontaneously attacked by hydroxide (OH⁻) ions. As the attack progresses, a self-reconstruction reaction begins almost instantaneously, resulting in the growth of metal hydroxide (M–OH) on the M–G surface, along with rapid discoloration of the dispersion solution. Benefiting from the simultaneous attacking and reconstructing, the surface of M–G is composed of M–OH and metal ions (M^x+) exposed by partially etched M–G, showing positively changed zeta potential (Fig. S1†). The surface activated M–G (denoted as M-GH) can quickly assemble with negatively charged MXene, hence forming M-GH@MXene with tunable network structures (Fig. S2†). The hierarchical magnetic nanocrystal@C@MXene is obtained by subsequent thermal annealing of M-GH@MXene in an inert atmosphere, which decomposes M-GH to produce magnetic nanocrystals orderly embedded in a spherical carbon matrix.

Fig. 1 Schematic illustration of the preparation of hierarchical magnetic nanocrystal@C@MXene.

The synthesis of hierarchical ZnFe₂O₄@C@MXene (denoted as ZFCM) is taken as a model system. Monodisperse ZnFe-G spheres (diameter size: ca. 400 nm) were obtained by a solvothermal method.³² After a facile water induction, the colour of the corresponding dispersion changed from yellow to orange (Fig. 2a), implying the conversion of ZnFe-G to activated ZnFe-GH. And the obtained ZnFe-GH exhibits a similar morphology to ZnFe-G (Fig. 2b and S4†). Ti₃C₂T_x MXene (lateral size: ca. 0.5–1.5 μm, Fig. 2c) is prepared by a wet chemistry method.³³ Subsequently, the slowly added MXene nanosheets were rapidly assembled with the ZnFe-GH spheres driven by electrostatic force (Fig. S3†). After sitting for a few minutes, the obtained hybrids quickly settled at the bottom of the container, resulting in a colourless supernatant (Fig. 2a), which confirmed the successful assembly of ZnFe-GH spheres and MXene.

Fig. 2 (a) Digital images of ZnFe-G, ZnFe-GH aqueous dispersions, and corresponding ZnFe-GH@MXene precipitates. (b) SEM images of ZnFe-GH. (c) TEM images of MXene. (d–g) SEM, TEM, and HRTEM images of ZnFe-GH@MXene-3. (h) FTIR spectra and (i) XRD patterns of ZnFe-G, ZnFe-GH, MXene, and ZnFe-GH@MXene. (j) C 1s core-level XPS spectra of ZnFe-G and ZnFe-GH.

The morphology of ZnFe-GH@MXene-3 was observed with a scanning electron microscope (SEM) and transmission electron microscope (TEM). From Fig. 2d and e, the ZnFe-GH spheres are perfectly assembled with few MXene nanosheets due to the excellent flexibility of MXene. Simultaneously, the edges of the assembled nanosheets are self-folded through van der Waals forces, hydrogen bonds, and the centrifugal force generated by stirring (Fig. 2f), resulting in petal-like MXene shells that can provide local conductive networks anchored on the spherical surface.^9,34 In addition, high-resolution TEM (HRTEM, Fig. 2g) verifies that the MXene structure remains stable after assembly, showing periodic lattice fringes with an interlayer distance of 0.26 nm corresponding to the (100) plane of Ti₃C₂T_x MXene.³⁵Fig. 2h shows the Fourier transform infrared spectroscopy (FTIR) spectra of water-modified ZnFe-GH and self-assembled ZnFe-GH@MXene. The weakening of CH₂/CH₃ and C–O vibrations is observed in the FTIR spectrum of ZnFe-GH, indicating the partial removal of glycerate ligands.³⁶ And the pronounced peaks ascribed to Ti–O, C–H, C–O, and C=O vibrations are observed in the FTIR spectrum of ZnFe-GH@MXene, further confirming the synthesis of ZnFe-GH@MXene.⁸

X-ray photoelectron spectroscopy (XPS) and X-ray diffraction (XRD) were conducted to investigate the mechanism of the water-induced self-assembly method. The C 1s core-level XPS spectra of ZnFe-GH show a lower proportion of C–O and C=O relative to ZnFe-G (Fig. 2j), due to the surface deconstruction caused by the attack of OH⁻ ions in water. In the XRD pattern of ZnFe-G, there is only an obvious diffraction peak of about 12°, which is the characteristic of metal glycerate (Fig. 2i).³⁷ After water dispersion, the characteristic peak transformed into a board peak. And the corresponding partial enlargement of the XRD pattern of ZnFe-GH displays two new peaks at 34° and 61°, suggesting that a trace amount of hydroxide was synthesized in situ.³⁸ Since such self-reconstruction is a kind of time-dependent reaction, only a small amount of M–OH is generated on the surface of M–G in a short time.³⁹ Notably, the characteristic peak (002) of MXene is nearly invisible in the XRD pattern of ZnFe-GH@MXene, illustrating that MXene nanosheets are effectively isolated by ZnFe-GH spheres, resulting in the expanding of layer spacing.²⁶

Through the fine-tuning of the MXene loading, the size of MXene networks can be precisely regulated from the nanoscale to the microscale. Different from the 2D self-stacking structure of the pristine MXene, the petal-like MXene shells of preliminary assembled hybrids provide 3D connection spaces for the interaction with neighboring nanosheets (Fig. S6†). After subsequent thermal annealing of ZnFe-GH@MXene, the hierarchical magnetic ZFCM is generated. As shown in Fig. 3, all the samples perfectly inherit the 3D network morphologies of the assembled precursors. The ZFCM-3 hybrids with petal-like shells show good dispersion (Fig. 3d and i), and the morphology of the exposed core is exactly the same as the morphology of ZnFe₂O₄@C (denoted as ZFC, Fig. 3h), which is synthesized by annealing ZnFe-GH in a similar method (Fig. S8†). Likewise, the XRD patterns of the ZFCM series are similar to those of ZFC (Fig. S7†), indicating that the ZnFe-GH cores of the assembled precursors were successfully transformed into ZFC. As the MXene content in the assembled precursors is increased to 5%, the obtained ZFC cores are fully encapsulated and interconnected by MXene nanosheets, resulting in many local interconnected networks (Fig. 3e and j). By continuously increasing the load of MXene, a microscale 3D network structure was obtained for ZFCM-7, which is due to the bridge connection between MXene nanosheets and nanoscale 3D networks (Fig. 3f and k). Accordingly, it can be concluded that the morphologies of the resulting conductive networks can be directly controlled by adjusting the MXene load of assembled precursors, thereby modifying the dielectric properties of ZFCM for meeting the requirements of attenuating microwaves effectively.

Fig. 3 Formation scheme for (a) the magnetic hybrids with a tunable morphology and (b) 3D MXene. SEM images of (c and h) ZFC, (d and i) ZFCM-3, (e and j) ZFCM-5, (f and k) ZFCM-7, and (g) 3D MXene. TEM images of (m) ZFC, (n and o) ZFCM-3, and (l and q) 3D MXene. (p) HRTEM image and (r) EDS mapping images of ZFCM-3.

In the TEM image of ZFCM-3 (Fig. 3n and o), the core of the hybrid exhibits a similar morphology to ZFC (Fig. 3m), which is consistent with the SEM results. In the HRTEM of the exposed core (Fig. 3p), the ultrafine nanocrystals with periodic lattice fringes are orderly separated by the carbon matrix, and the lattice space of 0.48 nm is attributed to the (111) plane of ZnFe₂O₄ (JCPDS no. 22-1012, d = 0.487 nm).⁴⁰ The uniform carbon matrix derived from organic residues and massive ZnFe₂O₄ nanocrystals together form continuous heterogeneous interfaces. It is further verified by the high-density distribution of nanocrystals in the core of ZFCM, as evidenced in the EDS mapping of ZFCM-3 and ZFC (Fig. 3r and S9†).

For comparison, 3D networks composed only of MXene (denoted as 3D MXene) were synthesized via a hard template method. Compared to ZFCM-7 with core-supported 3D networks, the obtained 3D MXene exhibits obvious collapse and local aggregation, which would lead to undesired local impedance mismatch (Fig. 3l, q and S11†).¹⁹

As shown in Fig. 4a, the ZFCM series show good magnetism due to the isolated magnetic nanocrystals. From the hysteresis loops of the obtained samples, ZFC exhibits the strongest static magnetic properties (Fig. 4a). With the increase of MXene load, the saturation magnetization value and coercivity of the ZFCM series decreased, indicating that the structural adjustment of ZFCM affects the static magnetic properties of the resulting hybrids.

Fig. 4 (a) Magnetic hysteresis loops of ZFC and ZFCM series, and the inset shows their corresponding amplified views of the hysteresis loops at low applied fields. (b) Raman spectra of ZFC, MXene and ZFCM-7. (c) XPS survey spectra of ZFC and ZFCM-7. (d) C 1s core-level XPS spectra of ZFC and ZFCM-5. (e) Zn 2p and Fe 2p core-level XPS spectra of ZFCM-7. (f) Ti 2p core-level XPS spectra of MXene and ZFCM-7.

Fig. 4b shows the Raman spectrum of ZFC, MXene, and ZFCM-7. ZFCM-7 presents peaks centered at 198, 340, 398, 511, 575, 670, 1098, 1347, and 1580 cm⁻¹. The peaks at 198, 398, and 575 cm⁻¹ come from the contribution of MXene.⁴¹ The peaks at 340, 511, 670, and 1098 cm⁻¹ are assigned to the contribution of ZnFe₂O₄. Besides, the two broad bands at 1347 and 1580 cm⁻¹ are the characteristic D and G bands of carbon, which come from the carbon matrix of ZFC and the carbon exposed by mild oxidation of MXene^42,43.

X-ray photoelectron spectroscopy (XPS, Fig. 4c) reveals the coexistence of C, Ti, F, Fe, Zn, and O elements in ZFCM-7. Fig. 4d shows the C 1s core-level XPS spectrum of ZFCM-7, which could be fitted with five components ascribed to C–Ti (282.1 eV), C–Ti–T_x (283.5 eV), C–C (284.8 eV), C–O (286.1 eV), and C=O (288.9 eV), respectively.⁴⁴ The emergence of C–Ti and C–Ti–T_x relative to ZFC further validates the successful synthesis of ZFCM-7. The Fe 2p core-level XPS spectrum (Fig. 4e) of ZFCM-7 could be fitted with two components ascribed to Fe²⁺ (711.1 and 724.5 eV) and Fe³⁺ (714.0 and 727.8 eV). The Zn 2p core-level XPS spectrum could be fitted with one component ascribed to Zn²⁺ (1022.1 and 1045.2 eV).⁴⁵ The Ti 2p core-level XPS spectrum (Fig. 4f) of ZFCM-7 could be fitted with four components ascribed to Ti–C (455.1 eV), Ti²⁺ (456.0 eV), Ti³⁺ (457.0 eV), and Ti–O (458.9 eV), respectively.⁸ The oxidation of Ti atoms on the surface accounts for a higher proportion of Ti–O relative to the pristine Ti₃C₂T_x MXene, which could be ascribed to the mild oxidation of MXene by its oxygen-containing functional groups and the decomposition products of the assembled organometallic precursor.³⁴

To further validate the generality of the synthetic strategy, a series of hierarchical magnetic nanocrystal@C@MXene hybrids, including Fe₃O₄@C@MXene, CoO@C@MXene, and CuCo@C@MXene, were synthesized simply by using the respective precursors. All samples show strong magnetism under an external magnetic field (Fig. S2†), which is consistent with the VSM results (Fig. S12†). The formation of hybrids is confirmed by the corresponding XPS spectra and XRD patterns (Fig. S13 and S15†). The SEM images of the obtained magnetic hybrids exhibit a similar spherical morphology with an obvious petal-like MXene shell. TEM images provide direct evidence that the hierarchical structures are formed, while there is a slight difference in the morphological details of magnetic nanocrystals, which is due to the variations in the doping atoms of precursors (Fig. S14 and S15†). Moreover, the corresponding hybrids with microscale 3D networks were easily synthesized by increasing the MXene load in the assembled precursors (Fig. S16†), realizing the transformation from a nanoscale 3D network to a microscale 3D network. Our preliminary results suggest that the present approach is simple and general, and can be applied to synthesize hierarchical magnetic hybrids with tunable conductive networks and adjustable element doping.

To evaluate the MA performances of hierarchical structures with tunable networks, the electromagnetic parameters of samples were analyzed in the range of frequencies 2–18 GHz. In general, the real part of complex permittivity (ε′) and complex permeability (μ′) reflects the ability of absorbers to store electromagnetic energy, while the imaginary part of complex permittivity (ε′′) and complex permeability (μ′′) represents the loss capability of energy.⁴⁶ As shown in Fig. 5a and b, due to the introduction of MXene networks, both the ε′ and ε′′′ of ZFCM series are significantly elevated. The ε′ plots of all samples showed a decreasing trend with increasing frequency in the range of 2–18 GHz, which is consistent with Debye theory.⁴⁷ According to Debye theory, the conduction part and the polarization part can be separated from ε′′ and investigated as follows:^48,49

where τ represents the relaxation time, ε_s represents the static permittivity, ε_∞ represents the relative permittivity at an infinite frequency, ε₀ represents the vacuum dielectric constant, and σ represents the conductivity. Compared with low-conductivity ZFC, the ZFCM series exhibit significantly improved conductivity (Fig. S17†). With the increase of MXene load, the σ values of ZFCM series exponentially increase, further demonstrating the successful control of the assembled conductive networks by the fine-tuning of MXene load. Due to the positive correlation between and σ values, the calculated values of ZFCM series also increase with MXene load. However, the frequency-dependent values rapidly decrease with an increased frequency and are much smaller than in the range of about 4–18 GHz (Fig. S18†). Therefore, in the high frequency region, the main contribution to the dielectric loss of ZFCM series is polarization loss rather than conduction loss. The ε′′ and corresponding dielectric loss tangent (tan δ_ε = ε′′/ε′) plots of ZFCM series exhibited several relaxation peaks. And the relaxation peaks of ZFCM-7 are the most obvious, indicating the existence of prominent multiple polarization relaxation processes.⁵⁰ The Debye dipolar relaxation process can be further studied by using the Cole–Cole curve based on the following eqn (3):⁵¹

Fig. 5 (a–c) Real parts (ε′), imaginary parts (ε′′), and dielectric loss tangent values (tan δ_ε) of ZFC, ZFCM series, and 3D MXene. (d–f) real parts (μ′), imaginary parts (μ′′), and magnetic loss tangent values (tan δ_μ) of ZFC and ZFCM series. (g) Cole–Cole curves of ZFCM-7. (h) The calculated μ′′(μ′)⁻²f⁻¹ of ZFCM series. (i) Attenuation constants of ZFC, ZFCM series, and 3D MXene.

All ZFCM series exhibit undulant curves containing many semicircles (Fig. 5g and S19†), which may arise from interfacial polarization and dipolar polarization.^23,46 The continuous interfaces between the three phases of MXene, carbon matrixes, and magnetic nanocrystals can significantly trigger the interfacial polarization.⁵⁰ The residual end groups, the intrinsic defects, and the oxidation-generated defects of MXene as well as the defects in carbon matrixes can lead to the asymmetry of electron spatial distribution and the formation of dipole moment, resulting in dipolar polarization under microwave irradiation.^52–54

Due to the introduction of magnetic ZnFe₂O₄ nanocrystals isolated by the carbon matrix, multiple magnetic losses are introduced into all ZFCM series, showing magnetic spectrum curves (Fig. 5d and e).⁵⁵ The μ′ plots of ZFCM series exhibit higher μ′ in the high frequency range, representing its better capability of storing magnetic energy in this range.⁵⁶ Interestingly, the μ′ and μ′′ plots of ZFCM-7 have the most obvious fluctuation, probably attributed to the magnetic coupling effect of the assembled microscale 3D networks.⁵⁷ In the tested frequency range, the magnetic loss is mainly derived from magnetic resonance and eddy current loss. Normally, magnetic resonance can be divided into natural resonance and exchange resonance, which always takes place in the range of 0.1–10 GHz and over the 10 GHz range, respectively.⁵⁸ The magnetic loss tangent value (tan δ_μ = μ′′/μ′) plots of ZFCM series exhibit strong fluctuations in the range of 2–6 GHz, which can be recognized as a symbol of natural resonance.⁵⁹ And the fluctuations of tan δ_μ plots correspond to the magnetic exchange resonance in the range of 12–18 GHz. The magnetic eddy current loss can be evaluated by using μ′′(μ′)⁻²f⁻¹, where μ′ and μ′′ correspond to the real and imaginary parts of complex permeability, respectively, and f represents the electromagnetic field frequency. When eddy current loss exists, μ′′(μ′)⁻²f⁻¹ does not change with frequency. In Fig. 6a, all μ′′(μ′)⁻²f⁻¹ plots of ZFCM series remain constant at 9–18 GHz, revealing the existence of eddy current loss in the high frequency range (Fig. 5h).⁶⁰ Meanwhile, an attenuation constant (α) is introduced to describe the loss capacity of an absorber as described in eqn (4):⁵¹

Fig. 6 3D RL values of (a) ZnFe₂O₄, (b) ZFC, (c) ZFCM-3, (d) ZFCM-5, (e) ZFCM-7, and (f) 3D MXene. (g) 2D contour maps of the |Z_in/Z₀| values of ZFC, ZFCM series, and 3D MXene.

It can be concluded from the formula that the attenuation constant (α) increases with the dielectric loss and magnetic loss capacity. As shown in Fig. 5i, the α values of ZFCM series are significantly improved with the increase of MXene load, demonstrating that the loss capacity can be enhanced by regulating the size of MXene networks.¹¹ In addition, the 3D MXene composed entirely of the MXene exhibits the highest α value, which is due to its excellent dielectric properties.

To intuitively evaluate the electromagnetic absorbing capacity of materials, the reflection loss (RL) was calculated based on the transmission line equation as follows:

where Z_in and Z₀ are the input impedance of absorbers and the impedance of free space, respectively. f is the frequency of electromagnetic waves, d is the thickness of the absorbing layers, and c is the velocity of light in free space.⁶¹

The plots of RL versus frequency and thickness are shown in Fig. 6 and S21.† For comparison with multiphase hybrids, pristine ZnFe₂O₄ was prepared (Fig. S10†). The pristine ZnFe₂O₄ exhibits negligible absorption of electromagnetic waves (Fig. 6a) due to low dielectric properties and impedance mismatch (Fig. S20†). The ZFC with carbon matrixes displays an RL_min value of −16.38 dB with a thickness of 2.50 mm (Fig. 6b). On account of the highly efficient synergistic effect among the MXene shell, carbon matrixes, and magnetic nanocrystals, ZFCM-3 and ZFCM-5 display RL_min values of −57.66 dB at 7.24 GHz and −61.19 dB at 8.81 GHz, respectively (Fig. 6c and d). Notability, ZFCM-7 with a microscale 3D network structure represents the most prominent RL_min value of −62.59 dB at 14.16 GHz and an excellent EAB of 4.42 GHz, and the corresponding fitting thickness reaches an astonishing 1.33 mm, realizing high MA under ultrathin thickness (Fig. 6e). Conversely, the 3D MXene with an excellent attenuation constant only offers an unsatisfactory RL_min value of −7.94 GHz, which originates from the incompatibility between the high dielectric constant and impedance matching (Fig. 6f and S22†).¹¹

To further explore the effect of the tunable 3D structure on the impedance matching of ZFCM, the normalized characteristic impedance (|Z_in/Z₀|) is introduced, which can intuitively evaluate the degree of impedance matching. According to transmission line theory, when the value of |Z_in/Z₀| is close to or equal to 1, the impedance matching is favorable, which demonstrates that the electromagnetic wave can successfully enter the absorber instead of turning into an undesired reflection.⁶² In the 2D contour maps of |Z_in/Z₀| values (Fig. 6g), the deep blue region belongs to the region with good impedance matching. Different from the complete impedance mismatch of 3D MXene and ZnFe₂O₄ (Fig. S20b†), ZFC achieves impedance matching in a large area. However, the |Z_in/Z₀| value of ZFC changes rapidly with frequency, resulting in a relatively high RL_min and an imperfect EAB of 3.48 GHz. After being assembled with the MXene, the matching region of the obtained ZFCM moves towards higher frequency and lower thickness with the increase of MXene load. Notably, ZFCM-7 with microscale 3D networks exhibits a mostly continuous impedance matching region, concentrating in the ultrathin thickness and the high frequency region, which is consistent with the RL result. Moreover, the MA performance of CuCo@C@MXene with microscale 3D networks (denoted as CCCM-7) was also tested. Similar to the impedance performance of ZFCM-7, the impedance matching region of CCCM-7 is concentrated in the high frequency and low thickness region, resulting in a highly efficient MA under an ultrathin thickness of 1.40 mm, and the corresponding EAB reaches 5.60 GHz, further elucidating the superiority of the 3D hierarchical network structure (Fig. S23†).

Overall, based on the details (the RL_min values and the broadest EAB) of the MA properties exhibited by ZFCM with different morphologies, it is obvious to conclude that the facile synthesis strategy of 3D hierarchical structures can synergistically integrate the advantages of MXenes, carbon matrixes and magnetic nanocrystals to achieve a superior MA performance (Fig. 7a). Besides, the step-by-step regulation of the MXene networks realizes the thinning of absorbers (Fig. S24†). The special RL values (SRL = RL_min/thickness) and the special EAB (SEAB = EAB/thickness) are introduced to reasonably evaluate MA ability considering the thickness.^63,64 As shown in Fig. 7b and Tables S1 and S2,† compared to other MXene-based absorbers and 3D hierarchical absorbers, ZFCM-7 prepared in this work exhibits better wave absorption properties at a lower thickness, achieving a balance of both strong absorption and a wide EAB.

Fig. 7 (a) Comparison of the RL_min values and broadest EAB of ZFC, ZFCM series, and 3D MXene. (b) SRL and SEAB of MXene-based absorbers. (c) Complex electric field intensity distribution and magnetic loss distribution of 3D MXene and ZFCM-7. (d) Schematic illustration of the MA mechanism for hierarchical magnetic nanocrystal@C@MXene with a 3D network structure.

To further investigate the influence of the hierarchical magnetic structure on the electromagnetic wave loss paths and loss methods, the electric field intensity distribution and magnetic loss distribution of 3D MXenes and ZFCM-7 were simulated by the finite element method based on COMSOL Multiphysics. As shown in Fig. 7c, an inhomogeneous electric field is concentrated inside the cavity structure of 3D MXene. The relatively high local field strength performance is attributed to the reflection and scattering of electromagnetic waves inside the cavity, which reveals that some of the electromagnetic waves can penetrate the MXene layer and propagate to the interior of the hollow sphere. As for ZFCM-7 with multiple heterogeneous interfaces, electric fields with strong energy density are present at the massive interfaces and magnetic nanocrystals. This is attributed to the local electric field distortion caused by the dielectric constant difference among the multiple heterostructures, which can provide evidence for the existence of strong interfacial polarization.⁶⁵ In the overall view and cross-section view of simulated magnetic loss, compared with the negligible magnetic loss performance of 3D MXenes, ZFCM-7 exhibits a significant magnetic loss distribution. Notably, the magnetic loss exhibits an array-like dispersion with high energy density, which is derived from the high-density carbon-encapsulated magnetic nanocrystals. At the same time, it can be found that strong magnetic loss distribution is also intense in the MXene shell and the carbon matrix, which might be attributed to the magnetic coupling between neighboring encapsulated magnetic nanocrystals.¹⁹ The heterojunction composed of highly separated magnetic nanocrystals and the carbon matrix effectively suppresses the skin effect, enhances their magnetic interaction, and induces the formation of high-density magnetic coupling networks, which can further promote the magnetic loss.⁶⁶ Hence, the construction of the 3D hierarchical magnetic structure successfully introduced various loss modes such as interface polarization and magnetic coupling, which effectively extended the microwave absorption path.

Accordingly, with the results of the above experiments and simulation analysis, the absorbing mechanism of the 3D hierarchical magnetic structure is initially proposed as shown in Fig. 7d. First, benefiting from the water-induced self-assembly strategy, the original self-stacked structure of MXene is greatly destroyed. The surfaces of the obtained samples are composed of randomly oriented and loosely connected MXenes, which facilitates the efficient entry of electromagnetic waves rather than reflection. Simultaneously, the external MXene nanosheets formed a microscale 3D conductive network, which can enable the free migration of excited electrons generated by electromagnetic waves. Second, the hierarchical structure endows samples with continuous heterointerfaces, resulting in a rich interface polarization loss between the three phases of MXene, carbon matrixes, and magnetic nanocrystals. The residual end groups on the MXene surfaces, as well as the defects of MXenes and carbon matrixes, can lead to the asymmetry of electron spatial distribution and the formation of dipolar polarization sites, providing dipole polarization losses. Third, the isolated magnetic nanocrystals with a high density provide multiple magnetic losses including magnetic resonance and eddy current loss, further eroding the energy of electromagnetic waves. Moreover, the folded MXene on the 3D shell promotes multilevel reflection and scattering, prolonging the path of attenuation. As a result, the unique hierarchical structure delivers excellent MA performance at ultrathin thickness through the synergistic high efficiency of high-density magnetic components and high-loss dielectric networks.

Conclusions

In conclusion, we developed a simple and general synthetic strategy for obtaining ZnFe₂O₄@C@MXene and other series of MXene-coated hierarchical magnetic hybrids through water-induced self-assembly and subsequent heat treatment. The involved novel water-induced self-assembly method can assemble 2D MXene nanosheets into 3D networks with a tunable size, which is extremely effective in preventing the self-accumulation of MXenes and providing controllable dielectric loss at the same time. The M-GH spheres, as the key part of the assembly, are transformed into high-density dispersed ultrafine magnetic nanocrystals with carbon matrixes after annealing, providing magnetic loss and dielectric loss. The obtained continuous heterointerfaces endow the hybrids with rich interface polarization among the three phases of MXene, carbon matrixes, and magnetic nanocrystals. Both experimental and theoretical studies reveal that the unique 3D hierarchical structure resulted in excellent MA properties, and all ZFCM series exhibit the RL_min values of about −60 dB. Particularly, through control over the MXene load, the size of 3D hierarchical networks can be regulated to precisely adjust the impedance matching. The obtained microscale 3D hierarchical networks show superior MA performance under ultrathin thicknesses. When the loading of MXene is fixed at 7 wt%, a microscale 3D hierarchical network structure is constructed, which endows the structure with superior MA performance under ultrathin thicknesses. ZFCM-7 exhibits a RL_min value of −62.59 dB and EAB of 4.42 GHz at a thickness of only 1.33 mm. Hence, the present work provides a flexible route for the structural design of 3D hierarchical network nanomaterials, which presumably facilitates the further development of ultrathin MA materials.

Author contributions

Jian Xu: conceptualization, data curation, formal analysis, investigation, methodology, and writing – original draft. Wenjun Ma and Peng He: formal analysis, methodology, validation, and visualization. Yukang Zhou: simulation analysis and investigation. Xiaoyun Liu: project administration, supervision, and writing – review & editing. Yi Chen: project administration, supervision, and writing – review & editing. Peiyuan Zuo: project administration, resources, and validation. Qixin Zhuang: funding acquisition, project administration, resources, and validation.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (52073091 and 2171086), Shanghai Pujiang Program (22PJ1402500), Natural Science Foundation of Shanghai (20ZR1414600) and Fundamental Research Funds for the Central Universities (JKD01221701).

Notes and references

1. A. Iqbal, F. Shahzad, K. Hantanasirisakul, M. K. Kim, J. Kwon, J. Hong, H. Kim, D. Kim, Y. Gogotsi and C. M. Koo, Science, 2020, 369, 446–450.
2. B. Wang, Q. Wu, Y. Fu and T. Liu, J. Mater. Sci. Technol., 2021, 86, 91–109.
3. Z. Wu, H. W. Cheng, C. Jin, B. Yang, C. Xu, K. Pei, H. Zhang, Z. Yang and R. Che, Adv. Mater., 2021, 34, 2107538.
4. G. Wang, S. J. H. Ong, Y. Zhao, Z. J. Xu and G. Ji, J. Mater. Chem. A, 2020, 8, 24368–24387.
5. Y. Zhang, Y. Yan, H. Qiu, Z. Ma, K. Ruan and J. Gu, J. Mater. Sci. Technol., 2022, 103, 42–49.
6. S. De, S. Acharya, S. Sahoo, J.-J. Shim and G. C. Nayak, Mater. Chem. Front., 2022, 6, 818–842.
7. X. Li, Z. Wu, W. You, L. Yang and R. Che, Nano-Micro Lett., 2022, 14, 73.
8. G. Wu, T. Li, Z. Wang, M. Li, B. Wang and A. Dong, Angew. Chem., Int. Ed., 2020, 59, 20628–20635.
9. R. Sun, H.-B. Zhang, J. Liu, X. Xie, R. Yang, Y. Li, S. Hong and Z.-Z. Yu, Adv. Funct. Mater., 2017, 27, 1702807.
10. X. Li, X. Yin, C. Song, M. Han, H. Xu, W. Duan, L. Cheng and L. Zhang, Adv. Funct. Mater., 2018, 28, 1803938.
11. H. Y. Wang, X. B. Sun, S. H. Yang, P. Y. Zhao, X. J. Zhang, G. S. Wang and Y. Huang, Nano-Micro Lett., 2021, 13, 206.
12. M. Chang, Z. Jia, S. He, J. Zhou, S. Zhang, M. Tian, B. Wang and G. Wu, Composites, Part B, 2021, 225, 109306.
13. F. Wu, Z. Liu, J. Wang, T. Shah, P. Liu, Q. Zhang and B. Zhang, Chem. Eng. J., 2021, 422, 130591.
14. Z. Wang, Z. Cheng, C. Fang, X. Hou and L. Xie, Composites, Part A, 2020, 136, 105956.
15. H.-Y. Wang, X.-b. Sun and G.-S. Wang, J. Mater. Chem. A, 2021, 9, 24571–24581.
16. X. M. Guan, Z. H. Yang, M. Zhou, L. Yang, R. Peymanfar, B. Aslibeiki and G. B. Ji, Small Struct., 2022, 3, 2200102.
17. S. Wang, D. Li, Y. Zhou and L. Jiang, ACS Nano, 2020, 14, 8634–8645.
18. L. Liang, Q. Li, X. Yan, Y. Feng, Y. Wang, H. B. Zhang, X. Zhou, C. Liu, C. Shen and X. Xie, ACS Nano, 2021, 15, 6622–6632.
19. C. Wen, X. Li, R. Zhang, C. Xu, W. You, Z. Liu, B. Zhao and R. Che, ACS Nano, 2022, 16, 1150–1159.
20. Z. Liu, D. Lu, W. Wang, L. Yue, J. Zhu, L. Zhao, H. Zheng, J. Wang and Y. Li, ACS Nano, 2022, 16, 4642–4653.
21. F. Song, G. Li, Y. Zhu, Z. Wu, X. Xie and N. Zhang, J. Mater. Chem. A, 2020, 8, 18538–18559.
22. X. Li, W. You, L. Wang, J. Liu, Z. Wu, K. Pei, Y. Li and R. Che, ACS Appl. Mater. Interfaces, 2019, 11, 44536–44544.
23. F. Chen, S. Zhang, B. Ma, Y. Xiong, H. Luo, Y. Cheng, X. Li, X. Wang and R. Gong, Chem. Eng. J., 2022, 431, 134007.
24. L. Cai, F. Pan, X. Zhu, Y. Dong, Y. Shi, Z. Xiang, J. Cheng, H. Jiang, Z. Shi and W. Lu, Chem. Eng. J., 2021, 434, 133865.
25. Y. Guo, D. Zhang, Y. Yang, Y. Wang, Z. Bai, P. K. Chu and Y. Luo, Nanoscale, 2021, 13, 4624–4633.
26. Z. Zou, Q. Wang, J. Yan, K. Zhu, K. Ye, G. Wang and D. Cao, ACS Nano, 2021, 15, 12140–12150.
27. Q. Song, F. Ye, L. Kong, Q. Shen, L. Han, L. Feng, G. Yu, Y. Pan and H. Li, Adv. Funct. Mater., 2020, 30, 2000475.
28. J. Li, D. Zhou, P.-J. Wang, C. Du, W.-F. Liu, J.-Z. Su, L.-X. Pang, M.-S. Cao and L.-B. Kong, Chem. Eng. J., 2021, 425, 131558.
29. B. Deng, Z. Liu, F. Pan, Z. Xiang, X. Zhang and W. Lu, J. Mater. Chem. A, 2021, 9, 3500–3510.
30. F. Hu, X. Wang, S. Bao, L. Song, S. Zhang, H. Niu, B. Fan, R. Zhang and H. Li, Chem. Eng. J., 2022, 440, 135855.
31. X. Li, C. Wen, L. Yang, R. Zhang, X. Li, Y. Li and R. Che, Carbon, 2021, 175, 509–518.
32. L. Shen, L. Yu, X. Y. Yu, X. Zhang and X. W. Lou, Angew. Chem., Int. Ed., 2015, 54, 1868–1872.
33. Y. Xia, T. S. Mathis, M. Q. Zhao, B. Anasori, A. Dang, Z. Zhou, H. Cho, Y. Gogotsi and S. Yang, Nature, 2018, 557, 409–412.
34. M. Q. Zhao, X. Xie, C. E. Ren, T. Makaryan, B. Anasori, G. Wang and Y. Gogotsi, Adv. Mater., 2017, 29, 1702410.
35. W. Eom, H. Shin, R. B. Ambade, S. H. Lee, K. H. Lee, D. J. Kang and T. H. Han, Nat. Commun., 2020, 11, 2825.
36. I. Saptiama, Y. V. Kaneti, Y. Suzuki, K. Tsuchiya, N. Fukumitsu, T. Sakae, J. Kim, Y. M. Kang, K. Ariga and Y. Yamauchi, Small, 2018, 14, 1800474.
37. L. Shen, L. Yu, H. B. Wu, X. Y. Yu, X. Zhang and X. W. Lou, Nat. Commun., 2015, 6, 6694.
38. J. Zhao, Y. Zou, X. Zou, T. Bai, Y. Liu, R. Gao, D. Wang and G. D. Li, Nanoscale, 2014, 6, 7255.
39. Y. V. Kaneti, R. R. Salunkhe, N. L. Wulan Septiani, C. Young, X. Jiang, Y.-B. He, Y.-M. Kang, Y. Sugahara and Y. Yamauchi, J. Mater. Chem. A, 2018, 6, 5971–5983.
40. J. Feng, Y. Hou, Y. Wang and L. Li, ACS Appl. Mater. Interfaces, 2017, 9, 14103–14111.
41. A. Sarycheva and Y. Gogotsi, Chem. Mater., 2020, 32, 3480–3488.
42. Y. Shi, Z. Xiang, L. Cai, F. Pan, Y. Dong, X. Zhu, J. Cheng, H. Jiang and W. Lu, ACS Nano, 2022, 16, 7816–7833.
43. L. Cabo-Fernandez, F. Mueller, S. Passerini and L. J. Hardwick, Chem. Commun., 2016, 52, 3970.
44. J. Halim, K. M. Cook, M. Naguib, P. Eklund, Y. Gogotsi, J. Rosen and M. W. Barsoum, Appl. Surf. Sci., 2016, 362, 406–417.
45. Z. Jiang, X. Han, Y. Du, Y. Li, Y. Li, J. Li, J. Tian and A. Wu, Nano Lett., 2021, 21, 2730–2737.
46. W. J. Ma, P. He, J. Xu, X. Y. Liu, S. L. Lin, Z. K. Cui, P. Y. Zuo and Q. X. Zhuang, J. Mater. Chem. A, 2022, 10, 11405–11413.
47. Q. Zhang, D. Xu, Y. Si, R. Xu, B. Luo, S. He and D. Liu, ACS Appl. Nano Mater., 2022, 5, 7175–7186.
48. Y. Li, W.-q. Cao, J. Yuan, D.-w. Wang and M.-s. Cao, J. Mater. Chem. C, 2015, 3, 9276–9282.
49. Z. Mu, G. Wei, H. Zhang, L. Gao, Y. Zhao, S. Tang and G. Ji, Nano Res., 2022, 15, 7731–7741.
50. J. Wang, L. Liu, S. Jiao, K. Ma, J. Lv and J. Yang, Adv. Funct. Mater., 2020, 30, 2002595.
51. S. Ren, H. Yu, L. Wang, Z. Huang, T. Lin, Y. Huang, J. Yang, Y. Hong and J. Liu, Nano-Micro Lett., 2022, 14, 68.
52. Y. Wu, Y. Zhao, M. Zhou, S. Tan, R. Peymanfar, B. Aslibeiki and G. Ji, Nano-Micro Lett., 2022, 14, 171.
53. J. Luo, M. Feng, Z. Dai, C. Jiang, W. Yao and N. Zhai, Nano Res., 2022, 15, 5781–5789.
54. X. Sang, Y. Xie, M. W. Lin, M. Alhabeb, K. L. Van Aken, Y. Gogotsi, P. R. C. Kent, K. Xiao and R. R. Unocic, ACS Nano, 2016, 10, 9193–9200.
55. J. Chen, J. Zheng, Q. Huang, F. Wang and G. Ji, ACS Appl. Mater. Interfaces, 2021, 13, 36182–36189.
56. X. Huan, H. Wang, W. Deng, J. Yan, K. Xu, H. Geng, X. Guo, X. Jia, J. Zhou and X. Yang, Small, 2022, 18, 2105411.
57. C. Zhang, Z. Wu, C. Xu, B. Yang, L. Wang, W. You and R. Che, Small, 2022, 18, 2104380.
58. Y. Liu, Z. Yao, J. Zhou, L. Jin, B. Wei and X. He, Carbon, 2022, 186, 574–588.
59. X.-L. Shi, M.-S. Cao, J. Yuan and X.-Y. Fang, Appl. Phys. Lett., 2009, 95, 163108.
60. B. Deng, Z. Xiang, J. Xiong, Z. Liu, L. Yu and W. Lu, Nano-Micro Lett., 2020, 12, 55.
61. Z. Wu, Z. Yang, C. Jin, Y. Zhao and R. Che, ACS Appl. Mater. Interfaces, 2021, 13, 5866–5876.
62. B. Fan, L. Xing, K. Yang, F. Zhou, Q. He, G. Tong and W. Wu, Chem. Eng. J., 2023, 451, 138492.
63. Y. Zhao, H. Zhang, X. Yang, H. Huang, G. Zhao, T. Cong, X. Zuo, Z. fan, S. Yang and L. Pan, Carbon, 2021, 171, 395–408.
64. Y. Li, X. Liu, X. Nie, W. Yang, Y. Wang, R. Yu and J. Shui, Adv. Funct. Mater., 2019, 29, 1807624.
65. K. Bi, M. Bi, Y. Hao, W. Luo, Z. Cai, X. Wang and Y. Huang, Nano Energy, 2018, 51, 513–523.
66. P. Liu, S. Gao, G. Zhang, Y. Huang, W. You and R. Che, Adv. Funct. Mater., 2021, 31, 2102812.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental sections; zeta potentials, digital images, SEM images, TEM images, elemental mapping, XPS spectra, HRTEM images, XRD patterns, electrical conductivity values, and microwave absorption performance data of related samples. Table of microwave performance comparison. See DOI: https://doi.org/10.1039/d2ta07718c

---