Skip to main content

Graphene-Assisted Synthesis of Fe4N with Enhanced Microwave Absorption Performance


Synergistic composite design of dielectric and magnetic materials is a highly attractive strategy to attenuate an electromagnetic wave. Herein, an Fe4N/N-doped reduced graphene oxide (N-rGO) composite is synthesized with a facile solvothermal method plus nitridation process. Notably, the rGO not only provides dielectric loss, but also reduces the nitride potential during the nitridation process and promotes the phase formation of Fe4N from Fe3O4. The electromagnetic wave absorption performance of Fe4N/N-rGO is thoroughly studied and reveals that designed Fe4N/N-rGO composite has a minimum reflection loss of −53.1 dB at 14.8 GHz with a thickness of 1.6 mm. Additionally, the reflection loss value of −41.63 dB is also achieved at a thinner matching thickness of 1.38 mm. The mechanism of the enhanced microwave absorption performance is proposed and is mainly due to the synergistic electromagnetic wave absorption with compositional advantages of rGO and Fe4N as well as a proper impedance match.

Graphical Abstract


With the rapid development of electronic technology, human health and delicate devices are under the potential threat of electromagnetic (EM) radiation.1,2 The acknowledged strategy is to transform an EM wave into thermal or other forms of energy by using absorption materials, which must have desirable properties with excellent microwave attenuation ability, wide effective absorption bandwidth (EAB), and they must be light weight and thin.3,4 To date, the loss mechanism of an EM wave includes dielectric loss and magnetic loss contributed by conductive and magnetic materials, respectively. The consensus is that single magnetic or dielectric material cannot achieve satisfactory absorption performance because they do not perform well on all bandwidths due to different working mechanisms. It is well accepted that integration of magnetic material and dielectric material into one design is a suitable strategy to achieve good EM materials using their synergistic effect on dielectric loss and magnetic loss.

Iron-based materials with well-recognized magnetic properties, excellent physical and chemical properties have been widely utilized in microwave absorption, electronic devices, batteries and electrocatalysis.5,6 In recent years, iron nitrides (FexN, X = 2, 3, 4) as magnetic components have attracted great attention because they have better oxidation resistance than other iron-based magnetic materials, larger magnetic saturation and higher Snoek limit than that of ferromagnetic metal oxides. In addition, high corrosion resistance is also a favorable characteristic of iron nitrides.7,8 Currently, many works on iron nitrides for microwave absorption have been reported, and enhanced EW absorption properties were demonstrated. For example, Cui et al. prepared an Fe3N/C composite by using organic amine as the substitution for the nitrogen and carbon source. Their sample showed a reflection loss (RL) value of −42.53 dB at 17.5 GHz with a thickness of 1.35 mm.9 In another of their works, core-shell Fe2N@N-doped carbon was prepared by nitriding the cubic precursor of Fe2O3@polydopamine (PDA). The best performance of −59.3 dB at 17.1 GHz with 1.55 mm was observed.5 However, the above composites cannot easily meet the requirements of high RL value and broad EAB at the same time, and are thus unfavorable for practical applications. Therefore, it is still challenging to develop iron nitrides with good comprehensive EM properties.

Compared with Fe2N and Fe3N, Fe4N exhibits a higher saturation magnetization (Ms) value and a greater magnetic loss, which makes it an optimal magnetic candidate.7,10 However, it is worth noting that the required nitrogen potentials for Fe2N, Fe3N and F4N are fairly close, with ~33%, ~25% and ~20% atomic N, respectively.8 It is rather tough to synthesize Fe4N because of its lowest nitrogen potential. If the nitrogen potential is not precisely controlled, raw materials cannot be nitrided completely and multiphase products will appear. For example, when Fe2O3@polyaniline (PANI) composite was maintained at 510 °C for 5 h in ammonia gas, a main phase of Fe4N plus minor Fe3N and Fe3O4 were obtained due to the incomplete nitridation reaction.11 Therefore, a cautious nitrogen potential-controlling operation ought to be carried out for the synthesis procedure of Fe4N or Fe4N-based composites.

Graphene is a typical ultra-light dielectric loss material, possessing good conductivity, large surface area, which is beneficial for absorbing EM waves. Meanwhile, the exposed functional groups and defects on the surface of graphene can produce localized state and polarization relaxation for attenuating EM waves.12 Thus, graphene is viewed as one of the best choices to combine with Fe4N.

In this work, Fe4N/N-rGO was obtained by calcining the Fe3O4/rGO composite under ammonia atmosphere. The complete phase transformation from Fe3O4 to Fe4N was successfully realized. The electromagnetic properties were measured and an excellent microwave absorption ability at both low and high frequency was detected. The effects of rGO content and calcination temperature on the phase type of iron nitrides are discussed. In addition, the mechanism of microwave absorption process is investigated.

Experimental Section

Preparation of Fe3O4/rGO Composites

The Fe3O4 particles anchored on rGO nanosheets were prepared via a mild solvothermal method according to our group’s previous work.13 Composites with various GO content (0–50 wt.%) were marked as Fe3O4, S1, S5, S10, S30 and S50, respectively.

Preparation of Iron Nitride/N-rGO Composites

Because of the good atomic diffusion and phase transformation at high temperature,14 solid-phase reaction method was adopted to synthesize iron nitride/N-rGO composites. Ammonia gas is commonly used in the nitridation process.15 Thus, the as-prepared samples including Fe3O4, S1-S50, were annealed at a heating rate of 5 °C/min with an NH3 flow of 150 mL/min and maintained at 500 °C, 600 °C and 700 °C for 4 h, respectively. The products derived from S50 were labeled as S500, S600 and S700 while the products originating from Fe3O4 were labeled as F-500, F-600 and F-700.

Characterization and Measurement

The crystal structures of the prepared samples were characterized by X-ray diffraction (XRD) on an X-pert powder diffractometer with Cu Kα radiation (λ = 0.154056 nm). The morphology and microstructure of the samples were observed by scanning electron microscopy (SEM, ZEISS G300), transmission electron microscopy (TEM, JEOL JEM-2100). Energy dispersive X-ray spectroscopy (EDS) was used to analyze the elemental compositions of the samples. Raman spectra were obtained on a LabRAM HR Evolution spectrometer with an excitation laser wavelength of 532 nm. The magnetization hysteresis curves were measured on a MPMS-XL-5 vibration sample magnetometer with the applied field up to 10 kOe. X-ray photoelectron spectroscopy (XPS) measurement were carried out on Thermo-Scientific K-Alpha instrument (Al Kα source, 1486.6 eV). The electric conductivities of the samples were measured via a four-point probe (RTS-9) at room temperature.

The EM parameters were measured via a vector network analyzer (Agilent PNA-E8363B) in the frequency range of 2–18 GHz based on the coaxial method. The sample for microwave absorption measurement was mixed with a paraffin matrix at a loading ratio of 50 wt.% and then pressed into a toroid shape (φout, 7.00 mm; φin, 3 mm) with a thickness of 2 mm.

Results and Discussion

Figure 1a illustrates the synthesis process of S500, S600 and S700. Fe3O4/rGO samples were firstly obtained by the in situ growth of Fe3O4 particles on reduced graphene oxide nanosheets through a solvothermal method and then transferred into a tube furnace for a nitrogen treatment at high temperatures under an ammonia gas atmosphere. To clearly describe the state of the surface atoms, the schematic illustration for the nitridation process is shown in Fig. 1b. At the beginning, gaseous NH3 was captured by surface Fe atoms. Next, the N-H bond tended to be broken at a certain temperature and the generated movable H atoms combined with surface O atoms to form H2O. The oxygen vacancies formed after dehydration. As a result, the oxygen concentration difference between the inner material and surface was formed, which provided a driving force for the internal O atoms diffusion to the outside. Similarly, as the reaction proceeded, there was also a nitrogen concentration difference, which promoted the diffusion of surface N into the inside part of Fe3O4. The equilibrium concentration of Fe-N on the surface of the material determined the final product. Notably, the rGO consumed part of the ammonia gas for self-reduction, alleviating the nitrogen potential acting on the surface of Fe3O4, which was beneficial for the formation of Fe4N.

Fig. 1
figure 1

(a) The scheme of the synthesis process for S500, S600 and S700, (b) the schematic explanation for the nitridation process.

The XRD patterns of S50, S500, S600 and S700 are displayed in Fig. 2a. In our previous work, S50 is confirmed to be composed of Fe3O4 and rGO1313. Various crystal structures of the products are formed under different temperatures. To be specific, S500 has a similar spectrum to S50, except for the part marked with the black dotted box, in which the diffraction peaks located at about 38º, 40.8º, and 43.4º could be assigned to the (110), (002), and (111) lattice planes of Fe3N (JCPDS No. 83-0876). For S600, three distinct peaks at 41.1º, 47.9º and 70.1º match well with the lattice planes (111), (200) and (220) for Fe4N (JCPDS No. 83-0875). According to Scherrer's equation,\(D=\frac{0.89\lambda }{\beta cos\theta }\), where \(\lambda \) is the X-ray wavelength (0.154056 nm), \(\theta \) is the diffraction angle, \(\beta \) is the half width of the observed diffraction peak. The calculated particle size of S600 is 22 nm. For S700, apart from the peaks of Fe4N, other diffraction peaks located at about 38.2º, 41.1º, 43.6º and 44.6º correspond to the lattice planes of (110) (002) (111) for Fe3N (JCPDS No. 83-0877) and (110) for Fe (JCPDS No. 87-0821), respectively. The diffraction peaks of rGO at around 26° for S600 and S700 are invisible, which may be attributed to the weak crystallinity and scattering power.16 It can be concluded that the temperature plays a crucial role in affecting the formation of the novel Fe4N/rGO composites. Figure 2b illustrates the phase structures of F-500-700. The result shows that pure Fe3N is obtained when the temperature is 500 °C and mixed Fe3N/Fe4N is formed when the temperature increases to 600 °C and 700 °C. It suggests that the absence of rGO makes it difficult to synthesize pure Fe4N.

Fig. 2
figure 2

XRD patterns of (a) S50, S500, S600 and S700, (b) F-500, F-600 and F-700, (c) the lattice parameters of iron oxide and iron nitride, (d) Raman spectra of S50 and S600.

In order to shed light on the effect of rGO content on phase composition, S1-S30 was also treated with the same nitridation process and the corresponding XRD patterns are shown in supplementary Figure S1. Unfortunately, all products contain multiphases of iron nitrides instead of pure Fe4N. The rGO shares a certain amount of ammonia in the nitridation process. The reduced nitrogen potential builds a favorable formation environment for Fe4N. The crystal structures of Fe3O4 and Fe4N, as well as the corresponding lattice parameters are presented in Fig. 2c. The phase transformation can be seen intuitively. Raman spectra of S50, and S600 are shown in Fig. 2d. There are two broad peaks located at ~1350 cm-1 and ~1590 cm-1 representing the D band from disordered carbon or defective graphitic structure and the G band from the highly ordered graphitic lattice, respectively.17,18,19 The inset in Fig. 2d exhibits the typical 2D peak of graphene at around 2680 cm-1, suggesting the presence of rGO.20 The intensity ratio of D and G bands (ID/IG) is widely utilized to evaluate the defect density in carbon materials.21 The ID/IG value of S600 (1.248) is higher than that of S50 (1.054), which suggests that the incorporation of N atoms generates defects or lattice distortion and the electronic polarization is promoted.16

XPS data were used to determine the element composition and chemical bonding state for nanomaterials.22 The survey spectrum in Fig. 3a demonstrates the presence of Fe, N, C, O in S600, where the oxygen comes from the oxygen-containing functional groups remaining in rGO or the oxygen in the air. The high-resolution XPS spectra for C 1s, N 1s and Fe 2p are shown in Fig. 3b-d, respectively. The N 1s spectrum of S600 in Fig. 3c can be fitted into three peaks at 397 eV, 398.3 eV, and 400.3 eV, which are attributed to Fe-N, pyridinic-N, and pyrrolic-N, respectively.5,23 The three peaks at 284.7 eV, 285.5 eV, and 286.3 eV can be observed in the C 1s spectrum, as shown in Fig. 3b, and are assigned to C-C, C-N, and C=O bonds, respectively.24 The aforementioned analysis reveals that the C-N bonds originate from pyridinic or pyrrolic N bonding configuration, validating N-doping behavior in the rGO lattice.16 In the high-resolution Fe 2p spectrum in Fig. 3d, three doublet peaks are observed. The first doublet peaks located at 711.1 eV and 724.5 eV correspond to Fe 2p3/2 and Fe 2p1/2 split orbitals of Fe2+ ions9 while the second at 714.2 eV and 727.7 eV correspond to Fe 2p3/2 and Fe 2p1/2 split orbitals of Fe3+ ions. The third at 719.3 eV and 732.7 eV belong to shakeup satellite peaks.24 In addition, the characteristic peak at 707.3 eV corresponds to the Fe-N bond.5,25 Therefore, N atoms are successfully migrated into the composites after nitridation treatment.

Fig. 3
figure 3

XPS spectra of S600: (a) survey spectrum, (b) high-resolution of C 1s spectrum, (c) N 1s spectrum, (d) Fe 2p spectrum.

The morphologies of the samples are characterized with SEM and TEM images. Figure 4a shows that S50 is composed of stacked layered structures with some nanoparticles on it. After the nitridation treatment, a connected network structure with enlarged nanoparticles appeared in S600 as shown in Fig. 4b. The SEM-EDS spectrum (Fig. 4c) proves the presence of C, N and Fe in S600. The actual measured atomic ratio of Fe to N is 3.84 calculated from the table in Fig. 4c, which is slightly lower than that of Fe4N. This result is due to the extra N content in N-rGO. The corresponding elemental mappings of S600 in Fig. 4d–g demonstrate the uniform distribution of the three elements. The TEM image for S600 is displayed in Fig. 4h. Fe4N is made of nonuniformed nanoparticles which may be ascribed to the self-agglomeration caused by calcination and intrinsic magnetic properties of Fe4N. The inset in Fig. 4h displays the size distribution of nanoparticles in S600. The average size of the nanoparticle is 18.5 nm, which is basically consistent with the particle size calculated according to the Scherrer formula. The HRTEM images in Fig. 4i–j delineate the inter-planar distance of 0.219 and 0.190 nm for the nanoparticles, which are in line with the (111) and (200) planes of Fe4N. The selected area electron diffraction pattern of S600 shows the clear diffraction rings of (111), (200) and (220) planes, which is consistent with XRD results (Fig. 2a). The above analysis further confirms the formation of Fe4N.

Fig. 4
figure 4

SEM images of (a) S50 and (b) S600; (c) SEM-EDS spectrum, (d–g) element mapping results, (h) TEM, (i–j) HRTEM images and (k) selected area electron diffraction of S600.

The measured magnetic hysteresis loops of S50 and S600 at room temperature are shown in Fig. 5. Owing to the presence of non-magnetic rGO, the sample S50 exhibits a decreased Ms value of 47.4 emu/g, which is much smaller than the theoretical value of Fe3O4 (92 emu/g).26 After the nitridation treatment at 600 ℃, an improved Ms value of 106.6 emu/g is obtained, which can be attributed to the phase transformation from Fe3O4 to Fe4N. As the inset shows in Fig. 5a, the coercivity (Hc) values of S50 and S600 are around 14.5 Oe and 112.8 Oe, and the corresponding remanent magnetizations (Mr) are 2.3 emu/g and 3.4 emu/g, respectively. Furthermore, the comparison of Ms values among S600 and other similar composites consisting of carbon and magnetic materials [5, 7, 9, 24, 26,27,28] is listed in Fig. 5b. The excellent magnetic properties of S600 is good for the enhancement of magnetic loss, thereby promoting the absorbing performance.

Fig. 5
figure 5

(a) Magnetic hysteresis loops of S50 and S600 at room temperature. The inset shows the expanded low-field hysteresis curves. (b) The comparison of Ms for S600 with other similar systems.

The microwave absorption performance of absorbers can be assessed by RL values, which are calculated from the measured relative complex permittivity (\( \varepsilon _{{\text{r}}} = \varepsilon ^{\prime} - j\varepsilon '' \)) and permeability (\(\mu_{r} = \mu^{\prime} - j\mu ^{\prime\prime}\)) according to the following formula:29

$$ {\text{RL}} = 20\;{\text{log}}{\mid }\left( {Z_{{{\text{in}}}} - Z_{0} } \right)/(Z_{{{\text{in}}}} + Z_{0} ){\mid } $$
$$ Z_{{{\text{in}}}} = Z_{0} \sqrt {\mu_{{\text{r}}} /\varepsilon_{{\text{r}}} } {\text{tan}}\;h\left[ {j\left( {2\pi fd/c} \right)\sqrt {\mu_{{\text{r}}} \varepsilon_{{\text{r}}} } } \right] $$

where \({Z}_{\mathrm{in}}\) is the normalized input impedance of the absorber, \({Z}_{0}\) is the impedance of air, \(f\) is the frequency of the EM wave, \(d\) is the absorber thickness, \(c\) is the light velocity of EM waves in free space. Generally, an RL value less than -10 dB means that 90% of the EM wave energy can be dissipated.30

Figure 6a–b displays the 2D RL values of S50 and S600 with various thicknesses. S50 in Fig. 6a shows a weak microwave absorption performance with the minimum RL (RLmin) value of −12.51 dB at 14.4 GHz with a thickness of 1.5 mm. The poor performance is mainly caused by the mismatched impendence. Remarkably, after the phase transformation, a distinctly enhanced absorbing property of S600 in Fig. 6b is exhibited. RL values exceeding −30 dB can be gained in a wide frequency range of 4–17 GHz by choosing an appropriate absorber layer thickness between 1.5 mm and 4.5 mm. Moreover, it exhibits enhanced EM absorption properties at both low- and high-frequency regions. In detail, the RL value reaches −47.5 dB at 7 GHz with a thickness of 3 mm and −42.4 dB at 16.1 GHz with a thickness of 1.5 mm. Additionally, the absorption peaks in Fig 6a-b shift toward low frequency from high frequency with the increasing thickness, which can be explained by the quarter-wavelength matching model.31 The comparisons of EAB and RL values between S50 and S600 are shown in Fig. 6e–f. S600 has a wider EAB than that of S50 under the same thickness and higher RL values (d>1 mm). To explicitly show the RLmin result of S600, the 3D plot of calculated RL values along with thickness and frequency is shown in Fig. 6c–d. The RLmin value of S600 is −53.1 dB with thickness of 1.6 mm and the corresponding EAB is 5.19 GHz (12.81 GHz-18 GHz). The black solid line in Fig. 6d represents the contour line of -10 dB. Even at a very thin thickness of 1.38 mm, S600 still shows excellent microwave absorption performance with an RLmin of −41.63 dB as shown in Fig. 6g. In Fig. 6h, S600 exhibits the widest EAB of 5.27 GHz at a thin thickness of only 1.61 mm, and the corresponding RL value is −39.56 dB. According to the commercial standards, the layer thickness should be less than 2 mm, especially when applied to certain flight equipment. Therefore, the as-prepared absorbers show great potential in practical applications.

Fig. 6
figure 6

2D representation of RL values at different thickness for (a) S50 and (b) S600, the corresponding effective absorption bandwidth (e) and the bar chart of the RL values (f), (c) 3D plots of calculated RL values for S600 with thickness from 1 to 5 mm versus frequency and thickness, (d) contour map of the RL values for S600, reflection loss of S600 with the specific thickness of (g) 1.38 mm and (h)1.61 mm.

To further evaluate the absorbing ability of S600, the comparison of the optimal microwave absorption performance in this work with those reported previously are summarized in Table I. Evidently, S600 exhibits wide EAB and high RLmin at the same time with a thin thickness among the listed composites, confirming the superior application prospect.

Table I Comparison of the optimal microwave absorption performance in this work with those of iron nitride or iron-nitride based materials reported previously

The complex permittivity and complex permeability are investigated in a frequency range 2-18 GHz to further reveal the possible microwave absorption mechanism of the samples. It is known that the real parts of complex permittivity and permeability (\(\varepsilon{^{\prime}}\) and \(\mu {^{\prime}}\)) represent the storage ability of the electric or magnetic energy, while the imaginary parts of complex permittivity and permeability (\(\varepsilon {^{\prime}}{^{\prime}}\) and \(\mu {^{\prime}}{^{\prime}}\)) stand for the electric energy dissipation and magnetic loss, respectively.36 As shown in Fig. 7a–b, the \(\varepsilon {^{\prime}}\) and \(\varepsilon {^{\prime}}{^{\prime}}\) curves of S50 and S600 have similar trends in that the values decrease gradually with the increasing frequency. S50 processes higher \(\varepsilon {^{\prime}}\) and \(\varepsilon {^{\prime}}{^{\prime}}\) values than that of S600 across the entire frequency bandwidth. The decreased \(\varepsilon {^{\prime}}\) and \(\varepsilon{^{\prime}}{^{\prime}}\) values for S600 represents the smaller electron storage energy and dielectric loss. The dielectric loss is induced by conductive loss and the polarization loss, where the polarization loss includes interfacial polarization and dipole polarization.37 The inset in Fig. 7a illustrates the electric conductivities of S50 and S600 (101 S/m and 163 S/m, respectively). The doping of N in rGO and the formation of Fe4N promote the increase of the conductivity. Although the conductivity of S600 is enhanced, the dielectric loss is still lower than that of S50. The reasons are speculated as follows. Both interfacial polarization and dipole polarization play important roles in attenuating electromagnetic wave. For S50, Fe3O4 nanoparticles are monodispersed with small size of 12.3 nm,13 which produces a great number of interfaces. Additionally, there are also plentiful functional groups after solvothermal treatment, while for S600, a small sintering phenomenon is observed, which leads to an enlarged Fe4N particles with an average size of 22.0 nm. Moreover, rGO is further reduced in the nitridation procedure. As a result, the interfaces and functional groups are dramatically decreased. Consequently, S600 shows a decreased \(\varepsilon {^{\prime}}{^{\prime}}\) value.

Fig. 7
figure 7

Electromagnetic parameters of S50 and S600: (a) real and (b) imaginary parts of complex permittivity, (d) real and (e) imaginary parts of complex permeability, (c) dielectric loss tangent (f) magnetic loss tangent values, (g,h) the Cole-Cole plots, (i) C0 curves for S50 and S600.

The dielectric loss mechanism can be explained by the Debye relaxation theory, and the relation between \(\varepsilon {^{\prime}}\) and \(\varepsilon {^{\prime}}{^{\prime}}\) is described as follows:34

$$ \left( {\varepsilon {^{\prime}}{-}\frac{{\varepsilon_{s} + \varepsilon_{\infty } }}{2}} \right)^{2} + \left( {\varepsilon^{\prime\prime}} \right)^{2} = \left( {\frac{{\varepsilon_{s} - \varepsilon_{\infty } }}{2}} \right)^{2} $$

Where \({\varepsilon }_{s}\) is the static permittivity, \({\varepsilon }_{\infty }\) is the dielectric constant at the high-frequency limit. The plot of \(\varepsilon {^{\prime}}\) versus \(\varepsilon{^{\prime}}{^{\prime}}\) is a semicircle (Cole-Cole semicircle) and each one represents a Debye relaxation process. The Cole-Cole plots for S50 and S600 are shown in Fig. 7g–h. There are two main semicircles and a long smooth tail in Fig. 7g, indicating that S50 possesses relaxation processes and conductive loss behaviors. The retained oxygen-containing groups and defects on the surface of rGO can act as polarization centers, generating electron polarization relaxation under the altering electromagnetic field. The numerous interfaces between the smaller nanoparticle of Fe3O4 and rGO dominate the interfacial polarization which is induced by charge distribution differences at the phase boundaries and grain boundaries,2 while for S600 in Fig. 7h, one obvious semicircle and a long tail are observed. The retained oxygen-containing groups decreased dramatically after the nitridation treatment, leading to a weak dipole polarization. In addition, a few irregular and small Cole-Cole semicircles are observed in the above two samples, implying that some weak polarization relaxation processes may have occurred in S50 and S600.

The real and imaginary parts of complex permeability as a function of frequency within 2-18 GHz for S50 and S600 are shown in Fig. 7d–e. Typically, the real permeability for S50 in Fig. 7d decreases from 1.15 to 0.9 at 4.3 GHz first and then increases to close to 1 with small fluctuations. Because of the Snoek limit for Fe3O4, the \(\mu{^{\prime}}\) value for S50 tends to 1, exhibiting a non-magnetic characteristic in the frequency range of 7–18 GHz. Because of the phase transformation from Fe3O4 to Fe4N, S600 has a higher \(\mu {^{\prime}}\) (Fig. 7d) and \(\mu{^{\prime}}{^{\prime}}\) (Fig. 7e) values compared to that of S50, which suggests the increased storage ability of magnetic energy and magnetic loss. In addition, the \(\mu {^{\prime}}{^{\prime}}\) value of S600 is relatively stable between 0.1 and 0.23, which implies that the magnetic loss works in the whole frequency range.

The magnetic loss generally originates from the domain wall, natural and exchange resonances, magnetic hysteresis and eddy current loss.38 However, the domain wall resonance occurs when the frequency is below 100 MHz, and the magnetic hysteresis just exists when a strong magnetic field is applied. Thus, the above two factors can be excluded. For eddy current loss, the formula \({C}_{0}={\mu }^{\mathrm{^{\prime}}\mathrm{^{\prime}}}{({\mu }^{\mathrm{^{\prime}}})}^{-2}{f}^{-1}\) can be used to determine whether the magnetic loss is the result of eddy current. The C0 value should be kept constant in the tested frequency on the condition that the magnetic loss only results from the eddy current effect.39 As shown in Fig. 7i, the C0 curves for S50 and S600 fluctuate before 15 GHz, which demonstrates that the eddy current has a weak effect on magnetic loss in the corresponding frequency region, while in the high frequency (15–18 GHz), the C0 curve almost keeps stable with a slight fluctuation, indicating that eddy current contributes to magnetic loss. According to previous reports that the natural resonance occurs at frequency of 2-10 GHz and exchange resonance takes place in the frequency range of 10–18 GHz.38 Therefore, the multiple peaks observed in the grey circle can be assigned to natural and exchange resonances, respectively.

The dielectric loss tangent \((\mathrm{tan}{\delta }_{e}={\varepsilon }^{{^{\prime}}{^{\prime}}}/\varepsilon {^{\prime}}\)) and magnetic loss tangents (\(\mathrm{tan}{\delta }_{\mu }={\mu }^{{^{\prime}}{^{\prime}}}/\mu {^{\prime}}\)) are calculated to further assess the loss ability as shown in Fig. 7c and f. The \(\mathrm{tan}{\delta }_{e}\) value for S50 is close to 0.6. However, the \(tan{\delta }_{\mu }\) of S50 decreases from 0.4 to 0 before 7 GHz, suggesting that the magnetic loss is gradually reduced. For S600, \(\mathrm{tan}{\delta }_{e}\) and \(\mathrm{tan}{\delta }_{\mu }\) vibrate along certain value, indicating that both dielectric and magnetic loss have prominent effects on the attenuation of EM wave.

Attenuation constant (\(\alpha \)) is adopted to evaluate the attenuation characteristics in the interior of microwave absorber. It can be calculated as follows [40]:

$$ \alpha = \frac{\surd 2\pi f}{c}\sqrt {\left( {\mu {^{\prime\prime}}\varepsilon {^{\prime\prime}} - \mu {^{\prime}}\varepsilon {^{\prime}}} \right) + \sqrt {\left( {\mu {^{\prime\prime}}\varepsilon {^{\prime\prime}} - \mu {^{\prime}}\varepsilon {^{\prime}}} \right)^{2} + \left( {\mu^{\prime}\varepsilon^{\prime\prime} + \mu {^{\prime\prime}}\varepsilon {^{\prime}}} \right)^{2} } } $$

Figure 8a gives the attenuation constant \(\alpha \) for S50 and S600 in the frequency range of 2-18 GHz. These two samples have gradually increased \(\alpha \) values as the frequency increased, suggesting the high attenuation capacity. Although S600 has a lower \(\alpha \) value than that of S50, it still exhibits much better microwave absorption performance. This phenomenon is related to another factor, namely the impedance matching characteristic.

Fig. 8
figure 8

(a) Attenuation constants and impedance matching |\({Z}_{in}/{Z}_{0}\)| values at different thickness for S50 (b) and S600 (c).

The EM matching coefficient (Z) can be calculated with the following equation:41

$$ Z = |{Z_{{{\text{in}}}} /Z_{0}| = |\sqrt {\frac{{\mu_{{\text{r}}} }}{{\varepsilon_{{\text{r}}} }}} {\text{tan}}\;h\left[ {j\left( {\frac{2\pi }{c}} \right)\sqrt {\varepsilon_{{\text{r}}} \mu_{{\text{r}}} } fd} \right]}|$$

where \({Z}_{in}\) is the input impedance, \({Z}_{0}\) is the impedance of free space. In theory, perfect impedance matching (Z=1) allows all incident EM wave to enter the interior of the microwave absorber. By contrast, a strong EM reflection would be happened with poor impedance matching when Z values are apart from 1. S50 has a poor impedance matching as shown in Fig. 8b. Its Z values are apart from 1 in the whole frequency range, which suggests that the incident microwaves are difficult to enter into S500, while for S600 (Fig. 8c), Z values with different coating thicknesses are closer to 1, exhibiting good impedance matching conditions. This means that the incident microwaves can easily access S600. The good impedance matching of S600 is the primary reason for its excellent absorption performance.

Based on the above discussion, the main mechanisms responsible for the prominent microwave absorption performance of S600 can be concluded as follows. The matched impedance characteristics guarantee that the majority of the incident microwaves can be easily infiltrated into the absorbers. Then, the entered microwaves are attenuated through multiple mechanisms. The conductive network of N-rGO is favorable for electron transportation, resulting in a suitable conductive loss. The existence of interfaces such as between Fe4N particles and N-rGO nanosheets generated charge accumulation, which facilitated the formation of rich interfacial polarization. The presence Fe4N particles induce the generation of multiple reflection and scattering processes. Moreover, the natural resonance, exchange resonance and eddy current loss contribute to the magnetic loss for microwave absorption.


In this work, an Fe4N/N-rGO composite with enhanced EM wave absorption properties derived from Fe3O4/rGO was successfully synthesized. The reduced nitride potential caused by rGO provides suitable phase transformation environment for Fe4N. The composite shows excellent microwave absorption properties that the RLmin reaches −53.1 dB at 14.8 GHz with a thickness of 1.6 mm and the widest EAB is 5.27 GHz at 1.61 mm. At an even thinner thickness of 1.38 mm, it still exhibits good absorption properties of −41.63 dB. The excellent performance is due to the matched impedance, the magnetic loss of Fe4N nanoparticles, interfacial polarization between the abundant interfaces, and the conduction loss of rGO nanosheets. This work provides a strategy to fabricate Fe4N/carbon composite, enriching the research of FexN/carbon absorbing material.


  1. Z. Li, H. Lin, S. Ding, H. Ling, T. Wang, Z. Miao, M. Zhang, A. Meng, and Q. Li, Carbon 167, 148 (2020)

    CAS  Article  Google Scholar 

  2. W. Zhou, C. Jiang, X. Duan, J. Song, Y. Yuan, and N. Chen, Carbohydr. Polym. 245, 116531 (2020)

    CAS  Article  Google Scholar 

  3. H. Liang, H. Xing, M. Qin, and H. Wu, Compos. Part A: Appl. Sci. Manuf. 135, 105959 (2020)

    CAS  Article  Google Scholar 

  4. S. Dong, P. Hu, X. Li, C. Hong, X. Zhang, and J. Han, Chem. Eng. J. 398, 125588 (2020)

    CAS  Article  Google Scholar 

  5. X. Cui, X. Liang, J. Chen, W. Gu, G. Ji, and Y. Du, Carbon 156, 49 (2020)

    CAS  Article  Google Scholar 

  6. L. Zhang, W. Cai, and N. Bao, Adv. Mater. 33, e2100745 (2021)

    Article  Google Scholar 

  7. Y. Hu, R. Jiang, J. Zhang, C. Zhang, and G. Cui, J. Mater. Sci. Technol. 34, 886 (2018)

    CAS  Article  Google Scholar 

  8. S. Bhattacharyya, J. Phys. Chem. C 119, 1601 (2015)

    CAS  Article  Google Scholar 

  9. X. Cui, X. Liang, W. Liu, W. Gu, G. Ji, and Y. Du, Chem. Eng. J. 381, 122589 (2020)

    CAS  Article  Google Scholar 

  10. M. Yu, Y. Xu, Q. Mao, F. Li, and C. Wang, J. Alloys Compd. 656, 362 (2016)

    CAS  Article  Google Scholar 

  11. L. Dai, S. Xie, M. Yu, and L. Ci, J. Magn. Magn. Mater. 466, 22 (2018)

    CAS  Article  Google Scholar 

  12. C. Wang, X. Han, P. Xu, X. Zhang, Y. Du, S. Hu, J. Wang, and X. Wang, Appl. Phys. Lett. 98, 072906 (2011)

    Article  Google Scholar 

  13. J. Ren, L. Zhang, and X. Fan, J. Electron. Mater. 50, 5057 (2021)

    CAS  Article  Google Scholar 

  14. Z. Zhao, Z. Xue, Q. Xiong, Y. Zhang, X. Hu, H. Chi, H. Qin, Y. Yuan, and H. Ni, Sustain Mater. Technol. 30, e00357 (2021)

    Google Scholar 

  15. L. Zhang, W. Cai, J. Ren, and Y. Tang, Mater. Today Energy 21, 100703 (2021)

    CAS  Article  Google Scholar 

  16. X. Wang, Y. Lu, T. Zhu, S. Chang, and W. Wang, Chem. Eng. J. 388, 124317 (2020)

    CAS  Article  Google Scholar 

  17. T. Huang, Z. Wu, Q. Yu, D. Tan, and L. Li, Chem. Eng. J. 359, 69 (2019)

    CAS  Article  Google Scholar 

  18. S. Shen, L. Huang, X. Tong, R. Zhou, Y. Zhong, Q. Xiong, L. Zhang, X. Wang, X. Xia, and J. Tu, Adv. Mater. 33, e2102796 (2021)

    Article  Google Scholar 

  19. L. Huang, S. Shen, Y. Zhong, Y. Zhang, L. Zhang, X. Wang, X. Xia, X. Tong, J. Zhou, and J. Tu, Adv. Mater. (2021).

    Article  Google Scholar 

  20. J. Tang, N. Liang, L. Wang, J. Li, G. Tian, D. Zhang, S. Feng, and H. Yue, Carbon 152, 575 (2019)

    CAS  Article  Google Scholar 

  21. F. Wang, X. Li, Z. Chen, W. Yu, K.P. Loh, B. Zhong, Y. Shi, and Q.-H. Xu, Chem. Eng. J. 405, 126676 (2021)

    CAS  Article  Google Scholar 

  22. X. Shan, Y. Zhong, L. Zhang, Y. Zhang, X. Xia, X. Wang, and J. Tu, J. Phys. Chem. C 125, 19060 (2021)

    Article  Google Scholar 

  23. Z. Wu, K. Pei, L. Xing, X. Yu, W. You, and R. Che, Adv. Funct. Mater. 29, 1901448 (2019)

    Article  Google Scholar 

  24. V.K. Rohith, P. Saravanan, K.T.R. Suresh, R. Radha, M. Balasubramaniam, and S. Balakumar, Nanoscale 10, 12018 (2018)

    Article  Google Scholar 

  25. X. Fan, F. Kong, A. Kong, A. Chen, Z. Zhou, Y. Shan, and A.C.S. Appl, Mater. Interfaces 9, 32840 (2017)

    CAS  Article  Google Scholar 

  26. D. Sun, Q. Zou, Y. Wang, Y. Wang, W. Jiang, and F. Li, Nanoscale 6, 6557 (2014)

    CAS  Article  Google Scholar 

  27. S. Wei, X. Wang, B. Zhang, M. Yu, Y. Zheng, Y. Wang, and J. Liu, Chem. Eng. J. 314, 477 (2017)

    CAS  Article  Google Scholar 

  28. N. Zhang, Y. Huang, and M. Wang, Compos. Part B Eng. 136, 135 (2018)

    CAS  Article  Google Scholar 

  29. H. Wei, Y. Tian, Q. Chen, D. Estevez, P. Xu, H.-X. Peng, and F. Qin, Chem. Eng. J. 405, 126637 (2021)

    CAS  Article  Google Scholar 

  30. L. Wang, X. Yu, X. Li, J. Zhang, M. Wang, and R. Che, Chem. Eng. J. 383, 123099 (2020)

    CAS  Article  Google Scholar 

  31. Y. Zhang, Z. Yang, M. Li, L. Yang, J. Liu, Y. Ha, and R. Wu, Chem. Eng. J. 382, 123039 (2020)

    CAS  Article  Google Scholar 

  32. Q. Zheng, M. Yu, X. Gao, Y. Wang, Z. Zhang, H. Zhou, and Y. Dai, J. Mater. Sci 55, 16954 (2020)

    Article  Google Scholar 

  33. Q. Zheng, M. Yu, W. Wang, S. Liu, X. Liang, C. Wang, Y. Dai, and Y. Xu, Ceram. Int. 47, 8315 (2021)

    CAS  Article  Google Scholar 

  34. X. Zhou, B. Wang, Z. Jia, X. Zhang, X. Liu, K. Wang, B. Xu, and G. Wu, J. Colloid. Interf. Sci. 582, 515 (2021)

    CAS  Article  Google Scholar 

  35. M. Gong, C. Liu, J. Gao, A. Du, W. Tong, and C. Liu, J. Mater. Res. Technol. 9, 8646 (2020)

    CAS  Article  Google Scholar 

  36. M. Qiao, X. Lei, Y. Ma, L. Tian, X. He, K. Su, and Q. Zhang, Nano Res. 11, 1500 (2018)

    CAS  Article  Google Scholar 

  37. Y. Zhao, H. Zhang, X. Yang, H. Huang, G. Zhao, T. Cong, X. Zuo, Z. Fan, S. Yang, and L. Pan, Carbon 171, 395 (2021)

    CAS  Article  Google Scholar 

  38. C. Zhou, X. Wang, H. Luo, L. Deng, S. Wei, Y. Zheng, Q. Jia, and J. Liu, Chem. Eng. J. 383, 123095 (2020)

    CAS  Article  Google Scholar 

  39. J. Li, Y. Xie, W. Lu, and T.-W. Chou, Carbon 129, 76 (2018)

    CAS  Article  Google Scholar 

  40. X. Zhang, G. Ji, W. Liu, B. Quan, X. Liang, C. Shang, Y. Cheng, and Y. Du, Nanoscale 7, 12932 (2015)

    CAS  Article  Google Scholar 

  41. Y. Cheng, J. Cao, Y. Li, Z. Li, H. Zhao, G. Ji, Y. Du, and A.C.S. Sustain, Chem. Eng. 6, 1427 (2017)

    Google Scholar 

Download references


This work was supported by the National Key Research and Development Program of China (Grant No. 2020YFE0100100), the National Natural Science Foundation of China (Grant No. 51801180), the Key Research and Development Program of Jiangsu Province (Grant No. BE2018008-1) and the Key Research and Development Program of Sichuan Province (Scientific and Technological Cooperation of Sichuan Province with Institutes and Universities) (Grant No. 2020YFSY0001).

Author information

Authors and Affiliations


Corresponding author

Correspondence to Lingjie Zhang.

Ethics declarations

Conflict of interest

The authors declare that they have no conflict of interest.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary file1 (PDF 214 kb)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Ren, J., Zhang, L., Bao, N. et al. Graphene-Assisted Synthesis of Fe4N with Enhanced Microwave Absorption Performance. J. Electron. Mater. 51, 966–977 (2022).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:


  • Fe4N/N-rGO composite
  • nitridation process
  • nitride potential
  • microwave absorption