Solid and macroporous Fe3C/N-C nanofibers with enhanced electromagnetic wave absorbability

A series of solid and macroporous N-doped carbon nanofibers composed of Fe3C nanoparticles (named as solid Fe3C/N-C NFs, solid Fe3C/N-C NFs-1, solid Fe3C/N-C NFs-2, macroporous Fe3C/N-C NFs, macroporous Fe3C/N-C NFs-1 and macroporous Fe3C/N-C NFs-2, respectively) were prepared through carbonization of as-electrospun nanofiber precursors. The results show that the magnetic Fe3C nanoparticles (NPs) dispersed homogeneously on the N-doped carbon fibers; as-prepared six materials exhibit excellent microwave absorption with a lower filler content in comparison with other magnetic carbon hybrid nanocomposites in related literatures. Particularly, the solid Fe3C/N-C NFs have an optimal reflection loss value (RL) of −33.4 dB at 7.6 GHz. For solid Fe3C/N-C NFs-2, the effective absorption bandwidth (EAB) at RL value below −10 dB can be up to 6.2 GHz at 2 mm. The macroporous Fe3C/N-C NFs have a broadband absorption area of 4.8 GHz at 3 mm. The EAB can be obtained in the 3.6–18.0 GHz frequency for the thickness of absorber layer between 2 and 6 mm. These Fe3C–based nanocomposites can be promising as lightweight, effective and low-metal content microwave absorption materials in 1–18 GHz.

Following rapid development of electronic science and technology, widespread use of electronic devices in wireless communications, high-frequency circuit components and other related fields [1][2][3] . Considerable efforts have been put to prepare the effective electromagnetic (EM) samples due to its application in the military stealth technology [4][5][6] . In recent years, with the increasing attention on the prevention and control of electromagnetic pollution and the increasing requirements of military weapons, microwave absorbing samples with a low thickness, light density, high reflection loss at a wide frequency range and corresponding stealth technology are urgently needed [7][8][9] . And a number of studies have endeavored to the development of such materials 10,11 . For example, the flower-like CuS hollow materials were successfully prepared by a simple solvothermal procedure. This material can reach a lowest RL value of −31.5 dB with 30 wt% filler loading 3 . The Ni−CNT nanocomposites obtained the lowest RL values (−30.0 dB) at 2.0 mm thickness 10 . The CNFs/Fe absorbers exhibited super-duper EM wave absorbing ability on the basis of a lowest RL value (−67.5 dB) with merely 5 wt% filler loading 11 . Among these microwave-absorbing materials, carbon-based composite microwave-absorbing materials comprising of carbon and magnetic particles have been attracting increasing attention, including Fe 3 O 4 multi-walled carbon nanotube (CNTs) 9 , Fe 3 C/CNT nanocomposites 12 and porous Co/CNTs 13 . It is worth mentioning that our group prepared the cobalt/N-doped carbon nanocomposites with two morphologies using electrospinning and annealing methods 14 . The solid Co/N-doped carbon nanocomposites show a strong absorbing property of −24.5 dB at 3 mm with 10 wt% filler loadings. Because these materials not only have large EM losses, which can effectively improve EM wave absorption, but also can overcome the disadvantages of magnetic components with a high density [13][14][15][16][17] . The Fe 3 C nanoparticles possess satisfactory stability and outstanding magnetic properties [18][19][20] . Some studies have shown that the combination of Fe 3 C and carbon nanofibers can significantly improve the microwave absorption performance 21 . In order to meet the requirements of absorbing materials including light density, thin thickness, and high reflection loss at a wide frequency range, we have prepared the macroporous Fe 3 C/N-doped carbon nanofibers with a novel macroporous structure. The experimental results suggested that the macroporous Fe 3 C/N-doped carbon nanofibers were endowed with a lighter weight and a broader bandwidth contrasted to the  25 .
In this work, a series of solid and macroporous Fe 3 C/N-doped carbon hybrid nanocomposites have been obtained by carbonizing as-spun polyacrylonitrile (PAN)-based nanofiber precursors, which were synthesized by spinning a PAN/DMF (N, N-dimethyl formamide) solution containing iron acetylacetonate. The six nanocomposites with 10 wt% filler loading showed an exceedingly good microwave absorption property. The as-prepared lightweight magnetic carbon hybrid nanocomposites will be valuable in different EM absorbing areas due to their superior EM absorption performance. Figure 1 illustrates the preparation process of solid and macroporous Fe 3 C/N-doped carbon nanofibers by using electrospinning and calcination treatment. As shown in Fig. 1, the precursor nanofibers were fabricated using electrospinning method 11 . According to the literature 21,26,27 , in the heating process of the tube furnace, the polyacrylonitrile in the precursor fiber decomposes at 250-300 °C. The main decomposition products are hydrogen cyanide, ammonia and carbon, etc., however, the fiber morphology remains. Iron acetylacetonate starts to decompose into iron oxide at 280 °C. These Fe 3 C particles are distributed on the surface and inside of the nanofiber. As the temperature increases, the iron oxide reacts with carbon to form Fe 3 C nanoparticles, then the final sample of N-doped carbon nanofibers composed of Fe 3 C nanoparticles was produced 28,29 . The crystallographic features of the synthesized representative products were examined by XRD and Raman spectrometer ( 75-0910), respectively, revealing that the Fe 3 C nanoparticles and nanocrystalline graphite are successfully synthesized by the carbonization method 32,33 . In the Raman spectra depicted in Fig. 2, two broad peaks at 1342 cm −1 and 1580 cm −1 correspond to the D band and G band from carbon. As is known to all, the ratios of I D /I G indicate the carbon disorder degree 34,35 . In Fig. 2b, we calculated the value of I D /I G to be 1.08:1 for solid Fe 3 C/N-doped carbon nanocomposites, 1.05:1 for macroporous Fe 3 C/N-doped carbon nanocomposites and 1.03:1 for pure N-doped carbon nanofibers, respectively. It implies that the disorder degree of the carbon in our materials has greatly increased compared with pure N-doped carbon nanofibers. Moreover, the carbon with plenty of defects can be considered as effective polarization center under electromagnetic field 36 .

Results and Discussion
The SEM images (Fig. S1a,b) show the typical morphology of the as-obtained materials. The lengths of these two representative nanofibers could beyond several micrometers. The average diameters of solid Fe 3 C/N-C nanofibers are 500 nm, whereas the average diameters of macroporous Fe 3 C/N-C nanofibers are 1 μm (Fig. 3a,b). Their component contents are measured by EDS analysis (Fig. 3c,d). The results reveal the occurrence of C, N and Fe elements in as-synthesized two composites. For solid Fe 3 C/N-doped carbon nanofibers, the respective C, N and Fe mass fractions are 82.1%, 3.4% and 14.5%, while in the macroporous Fe 3 C/N-doped carbon nanofibers, the respective C, N and Fe mass fractions are 88.6%, 4.5% and 6.9%. The EDX mapping images and EDS spectra of solid Fe 3 C/N-doped carbon NFs-1, solid Fe 3 C/N-doped carbon NFs-2, macroporous Fe 3 C/N-doped carbon NFs-1 and macroporous Fe 3 C/N-doped carbon NFs-2 are shown in Fig. S2. For solid Fe 3 C/N-C NFs-1, the respective C, N and Fe mass fractions are 95.7%, 0.5% and 3.8%. As for solid Fe 3 C/N-C NFs-2, the respective C, N and Fe mass fractions are 80.6%, 4.4% and 15.0%. For macroporous Fe 3 C/N-C NFs-1, the respective C, N and Fe mass fractions are 90.0%, 4.8% and 5.2%. For macroporous Fe 3 C/N-C NFs-2, the respective C, N and Fe  Figure 4 shows typical TEM photographs and EDX mapping images. It is obvious that numerous Fe 3 C NPs are uniformly distributed in the two prepared materials (Fig. 4a,g). Further, the particle sizes of the Fe 3 C NPs ranged from 20 to 50 nm (Fig. 4b,h). Moreover, the Fe 3 C NPs in the solid Fe 3 C/N-doped carbon nanofibers present a lattice spacing of 0.221 nm between adjacent lattice, corresponding to the (120) lattice plane (Fig. 4c), indicating a high crystallinity. The elemental mappings ( Fig. 4d- Fig. 4g,h, the TEM and STEM images in Fig. S1c,d demonstrate that the macroporous Fe 3 C/N-doped carbon nanofibers have a pore size of around 500 nm. In Fig. 4i, the lattice distance is 0.300 nm, corresponding to the (111) diffraction peak of Fe 3 C 37 . The inset of Fig. 4i displays a selected area electron diffraction (SAED) pattern and the clear diffraction rings represent the (321) and (103) planes of the Fe 3 C nanoparticles. Moreover, the EDX mapping results confirm that Fe 3 C NPs are decorated in the macroporous Fe 3 C/N-doped carbon nanofibers (Fig. 4j-l). Figure 5 shows the survey spectra of the two representative nanocomposites. In the survey XPS spectra (Fig. 5a,b), the XPS peaks for C 1s, N 1s, and Fe 2p indicate the existence of these elements in these synthesized nanomaterials. Figure 5c,d represent the XPS spectra of C 1s for as-fabricated nanofibers. The C 1s spectrum includes three peaks that can be indexed to C-C at around 285.7 eV, C-N at around 284.7 eV and C-O at around 286.8 eV, respectively 38,39 . Figure 5e,f display N 1s spectra of the as-fabricated products. The N 1s spectrum include four different nitrogen atoms: pyridinic N (397.8 eV), pyrrolic N (399.4 eV), quaternary N (400.3 and 400.6 eV), and oxidized N (403.5 eV) 33 . The peaks at 708.5 eV is due to Fe 3 C. The other peaks shown in the spectra of Fe 2p in Fig. 5g could be divided into two pairs of peaks at around 710.4 and 724.4 eV, which represent Fe 2p 3/2 and Fe 2p 1/2 , respectively 40,41 . This confirms the existences of Fe 3 C in as obtained materials. As shown in Fig. 5h, in the macroporous Fe 3 C/N-C NFs, the peaks at 708.5 eV correspond to Fe 3 C, and the other two major peaks at around 710.6 and 725.0 eV represent the Fe 2p 3/2 and Fe 2p 1/2 , respectively.  Figure 6 illustrates the magnetic hysteresis loops of as synthetized Fe 3 C/N-C nanofibers materials measured at room temperature. From Fig. 6, it is not difficult to find that the two samples exhibit typical ferromagnetic properties at 300 K 42,43 . The saturation magnetization of the solid Fe 3 C/N-doped carbon nanocomposites reaches 14.8 emu/g and 9.4 emu/g for macroporous Fe 3 C/N-doped carbon nanocomposites. This phenomenon can be related to the higher iron contents for solid Fe 3 C/N-doped carbon nanocomposites 44 . As shown in Fig. 3c,d, the Fe mass fraction of the solid Fe 3 C/N-doped carbon hybrid nanocomposites (14.5%) is larger than that in macroporous Fe 3 C/N-doped carbon hybrid nanocomposites (6.9%). The specific surface area (SSA) of pure N-doped carbon nanofibers, solid and macroporous Fe 3 C/N-doped carbon nanofibers are 28.1, 144.3 and 30.6 m 2 g −1 , respectively (Fig. S3). As shown in Fig. S3, the two samples display IV N 2 adsorption isothermals, indicating that mesoporous structures in these materials. The SSA of the obtained materials was computed by Brunauer-Emmett-Teller method. Due to the limitations of the method, the 500 nm macropores is too large to be represented in the specific surface area data, so the macropores has no remarkable impact on the SSA of the macroporous Fe 3 C/N-doped carbon nanocomposites. The Fe 3 C/N-doped carbon hybrid nanocomposites have a larger SSA compared with the pure N-doped nanofibers, which indicates that the mass fraction of iron plays a major role in the specific surface area of these materials 45  nanofibers is smaller than the nanofibers with solid structure 47,48 , so the SSA of the macroporous nanocomposites is smaller than the solid nanocomposites. As is known to all, coercivity force is considered as a momentous parameter to evaluating the magnetic performance. The EM wave absorbers possess a high coercivity force value might result in a resonance effect 49 . In addition, the coercivity force value is closely bound up with the metal content. As a result, a high metal content may cause a high coercivity 50 . Herein, the solid nanofibers with the higher iron content are endowed with the higher coercive force of 421 Oe, whereas the coercivity force of macroporous nanofibers are 46 Oe. Figure 7a,b denote the curve of the real and imaginary parts (ε′ and ε″) of the complex permittivity as a function of frequency. For the two Fe 3 C/N-doped carbon hybrid nanocomposites, the ε′ values decrease from 14.6 to 9.12 and 12.6 to 5.2 in 1-18 GHz, respectively. The ε″ value of macroporous Fe 3 C/N-C NFs decreases from 6.0 to 3.2, while the ε″ values of solid Fe 3 C/N-C NFs increases from 3.65 to 5.43 and exhibits a peak in the 6.0-11.0 GHz ranges, which reflects a relatively high dielectric loss 51 . In contrast to the macroporous nanocomposites, the solid nanocomposites have a larger ε′ value, suggesting that higher appropriate conductivity is obtained. As shown in Fig. 7c,d, the real (μ′) and imaginary (μ″) part of the complex permeability for two samples remains constant, so we suppose that the complex permittivity may play an important role in the EM absorbing ability of the two nanofibers.
In order to investigate the main determinant for the as fabricate solid and macroporous nanofibers, we measured the dielectric dissipation factors tan d ε = ε″/ε′ and magnetic loss tangent tan d μ = μ″/μ′ of the two nanocomposites at 2 mm, respectively (Fig. 8). For the solid Fe 3 C/N-doped carbon nanofibers, the tan d ε and tan d μ curves have similar trends with the ε″ and μ″ curves (Fig. 7b,d), suggesting both dielectric loss and magnetic loss have an effect on the EM wave absorbing performance of the solid nanofibers, or, impedance matching 51,52 . Nevertheless, we didn't see similar results for macroporous Fe 3 C/N-doped carbon nanofibers. On the other hand, we also calculated Z = |Z in /Z 0 | of as-prepared samples as shown in Fig. 9. Figure 9a displays the normalized characteristic impedance of three as-prepared solid Fe 3 C/N-doped carbon nanofibers at 2.0 mm. The experimental results show that the Z values of the solid Fe 3 C/N-C NFs and solid Fe 3 C/N-C NFs-2 are approximately 1.0 (The nearer to 1.0 for Z, the better the impedance matching properties), suggesting the rather good impedance matching properties 53 . Figure 9b displays the impedance matching performance of three as-prepared macroporous Fe 3 C/N-doped carbon nanofibers at 2.0 mm. It is obvious that the macroporous Fe 3 C/N-C NFs have more appropriate impedance matching properties compared with the other two macroporous Fe 3 C/N-doped carbon nanocomposites. Besides, we also compare the impedance matching performance of the two representative Fe 3 C/N-C nanofibers and pure N-C nanofibers with the thickness of 2 mm and 3 mm (Fig. 9c,d). The Z values of solid Fe 3 C/N-C NFs are close to 1, which represents a remarkable impedance matching performance. Especially, for solid Fe 3 C/N-C nanofibers, the appropriate components, synergistic effect of conductive N-doped carbon nanofibers and higher composition of magnetic Fe 3 C nanoparticles are propitious to their outstanding impedance matching properties. The Z values of macroporous Fe 3 C/N-C NFs are far away from 1.0, indicating a relatively worse impedance matching properties. The Z values of pure N-C nanofibers are near to 0, revealing nearly none impedance matching properties. Thus we speculate that the solid Fe 3 C/N-C NFs have better impedance matching properties which can lead leading to the excellent microwave-wave absorbing ability.
The RL values were evaluated via measuring relative complex permittivity (ε r = ε′ − jε″) and permeability (μ r = μ′ − jμ″) based off of equations (1) and (2). The EM absorption property of solid Fe 3 C/N-doped carbon nanofibers with 10 wt% filler loading is shown in Fig. 10a. The respective optimal microwave absorption is achieved at 11.9 GHz for −29.1 dB (2 mm), 7.6 GHz for −33.4 dB (3 mm) and 5.4 GHz for −22.8 dB (4 mm), respectively. The EAB is 4.1 GHz (10.4-14.5 GHz) at 2.0 mm. In Fig. 10b, the macroporous Fe 3 C/N-doped carbon nanofibers with 10 wt% nanofiber loading shows an optimal RL value of −22.1 dB. The microwave absorbing bandwidths with RL lower than −10 dB are 4.8 GHz (8.2-13.0 GHz) at 3.0 mm. Apparently, the solid Fe 3 C/N-doped carbon nanocomposites present the optimal reflection loss value. Perhaps it is because the solid Fe 3 C/N-doped carbon nanofibers have the larger coercivity force and higher iron content 54,55 . Moreover, the proper impedance matching also plays a major role in its superior EM wave absorbing properties. As shown in Fig. 10c, the lowest RL value of the pure N-doped carbon nanofibers with 10 wt% doping amount is unable to achieve −10 dB at 2.0-6.0 mm. The EM absorbing ability of other two solid Fe 3 C/N-doped carbon nanofibers and two macroporous Fe 3 C/N-doped carbon nanofibers are shown in Fig. S4. In Fig. S4a, the optimal frequency value of solid Fe 3 C/N-C NFs-1 can reach −34.4 dB at 3.0 mm. It is worth mentioning that solid Fe 3 C/N-C NFs-2 with 6 mm possess an optimal RL min of −34.9 dB. The effective absorption bandwidth can reach 6.2 GHz across 11.8-18.0 GHz (Fig. S4c). The macroporous Fe 3 C/N-C NFs-1 displays the optimal EM absorption of −32.8 dB among the three macroporous Fe 3 C/N-doped carbon nanocomposites (Fig. S4b). The macroporous Fe 3 C/N-C NFs-2 displays the strongest RL value of −15.0 dB (Fig. S4d). Table 1 concludes the EM-wave absorbing ability of these Fe 3 C/N-doped carbon composite nanofibers. Table 2 shows that the present solid Fe 3 C/N-C NFs-2 and macroporous Fe 3 C/N-C NFs exhibits the outstanding EM wave absorbing ability among the related Fe 3 C-based materials with the smallest fiber loadings. For example, Fe-Fe 3 C/C microspheres with 25 wt% doping amount gained an optimal RL of −18.8 dB 12 . The Fe 3 C/graphitic carbon with 70 wt% doping amount loading achieved

Conclusions
In summary, solid and macroporous Fe 3 C/N-doped carbon nanocomposites have been successfully obtained by coupling electrospinning with heat treatment in an argon atmosphere. The nanocomposites possess superior microwave absorption capability at a lower doping content (10 wt%). The solid Fe 3 C/N-C NFs display the optimal RL value of −33.4 dB at 3 mm. The lower RL value can be attribute to the appropriate impedance matching, which can be achieved by changing the formation of the absorber and Fe 3 C content. The solid Fe 3 C/N-C NFs-2 present an optimal RL min of −34.9 dB with the widest EAB of 6.2 GHz. The minimum RL value of the macroporous was synthesized according to the description by Stöber et al. 11 . The macroporous Fe 3 C/N-doped carbon nanocomposites were prepared using the similar synthetic process with 1.2 g previously synthesized silica template in the spinning dope. The obtained precursor fibers were placed in sodium hydroxide solution and stirred for 2 h to remove the template, followed by washing with deionized water and desiccation. The sample gained was recorded as macroporous Fe 3 C/N-C NFs. Besides, the macroporous Fe 3 C/N-doped carbon nanofibers with the   Materials characterization. The three solid Fe 3 C/N-doped carbon nanofibers and three macroporous Fe 3 C/ N-doped carbon nanofibers have almost the same morphology, structure and magnetic performance, therefore the solid Fe 3 C/N-C NFs and macroporous Fe 3 C/N-C NFs are selected as the representative in each group to describe its morphology, microstructure and magnetic performance in details. The structure of the products was carried out by X-ray diffraction (XRD). Field-emission scanning electron microscopy (FESEM, Hitachi SU-8010), high-resolution transmission electron microscopy (HRTEM, JEM-2010), scanning transmission electron microscope (STEM), energy-dispersive X-ray spectroscopy (EDX) and energy-dispersive spectrometer (EDS) analysis were used to characterize the morphology, microstructures and element content of the representative solid and macroporous nanofibers. X-ray photoelectron spectroscopy (XPS) spectra were characterized by ESCALAB 250Xi spectrometer (Thermo Fisher). Raman spectra were measured by a Raman spectrometer. The hysteresis loops were achieved by superconducting quantum interference device (SQUID-MPMS3) at room temperature.

Measurement of electromagnetic parameters and absorbing performance.
The value of complex permittivity (ε′, ε″) and permeability (μ′, μ″) of the as-synthesized seven products were evaluated via a network analyzer (Model No. HP-E8362B, Agilent, frequency range 1.0 GHz to 18.0 GHz). The measurement samples were prepared by homogeneously milling the powders (10 wt%) in paraffin and then pressing them into a ring of 7.0 mm external diameter, 3.04 mm inner diameter, 2.0 mm thickness 59,60 . The EM-wave absorbing ability of the materials were computed by equations (1) and (2)