Figure 2a shows the XRD pattern of as-prepared Fe-bdc nanorods, and strong diffraction peaks, which match well with earlier reports, can be clearly seen [19]. Namely, highly crystallized pure Fe-bdc has been successfully synthesized. According to its SEM images (Fig. 2b, c), Fe-bdc powder is composed of spindle-like particles with a length of ~ 670 nm and a diameter of ~ 150 nm. After coated by TiO2 through hydrolysis method, the sample only shows three obvious diffraction peaks (Fig. 2d), which share the same location with Fe-bdc. Accordingly, TiO2 coating does not affect phase composition of Fe-bdc and TiO2 coating layer may be amorphous. As can be seen in Fig. 2e, f, spindle-like structure has been well preserved in Fe-bdc@TiO2 composite. Owing to the existence of granular TiO2 coating layer, rougher surface can be seen, and the length and width have increased to ~ 830 nm and ~ 210 nm, respectively. In order to select appropriate pyrolysis temperature in N2, TGA curve of Fe-bdc@TiO2 composite from 20 oC to 900 oC has been provided as Fig. 2g. Apparently, three major weight loss stages can be summarized as: (1) 20–332 oC: the weight loss (17.59 wt%) should be ascribed to the evaporation of solvent inside Fe-bdc and the complete polycondensation of polytitanoxane; (2) 332-502.1 oC: further weight loss (22.48 wt%) should be attributed to the decomposition of organic framework of Fe-bdc; (3) 502.1–685 oC: as for Fe-MOFs, the weight loss occurring at ~ 600 oC is always related with the carbothermal reduction of iron oxides [21]. In this case, the weight loss (10.33 wt%) is largely owing to the loss of O element and solid-state reaction between iron oxides and TiO2 under a reductive environment, which can be deduced by the XRD analysis. As can be seen in Fig. 2h, S-500 and S-600 have rather weak diffraction peaks, indicating low crystallinity. When the temperature increases to 700 oC, sharp diffraction peaks at 30.0o, 35.3o, 42.9o, 53.3o, 56.8o and 62.3o can be clearly observed. We should also mention that above three samples present similar diffraction peaks, which belong to (Fe2.5Ti0.5)1.04O4. When the temperature has increased to 800 oC, phase conversion takes place. And the peaks at 18.0o, 23.8o, 29.7o, 32.5o, 34.9o, 36.5o, 40.3o, 42.4o, 48.8o, 52.6o, 53.0o, 56.0o and 61.6o correspond well with characteristic diffraction peaks of Fe2TiO4 and FeTiO3. In brief, the solid-state reaction between iron-containing intermediate and TiO2 gives rise to (Fe2.5Ti0.5)1.04O4 at lower temperature, which would be further converted into Fe2TiO4 and FeTiO3 at higher temperature.
To further determine the element composition of FTO/C composites, Fig. 3a has summarized the ICP results. S-500, S-600, S-700, and S-800 own 47.84 wt%, 50.88 wt%, 65.2 wt% and 52.99 wt% of Fe, respectively. Thus, higher pyrolysis temperature leads to higher Fe content except for S-700. The Ti weight ratios of S-500, S-600, S-700 and S-800 are 20.48 wt%, 22.94 wt%, 24.18 wt% and 25.03 wt%, respectively, displaying a growing trend with rising temperature. In addition, the Fe/Ti molar ratios of S-500, S-600, S-700 and S-800 are 2.20, 1.90, 2.31, and 1.81, respectively. It is worth pointing out that the Fe/Ti molar ratios of S-500, S-600 and S-700 are less than theoretical value of (Fe2.5Ti0.5)1.04O4. Therefore, we may infer that there are residue TiO2 nanoparticles in above three samples. Raman spectra of FTO/C composites are further provided in Fig. 3b. Two peaks centered in ~ 1340 cm− 1 (D band) and ~ 1590 cm− 1 (G band) reveal the existence of partially graphitized carbon in all FTO/C composites [22]. Generally, the peak intensity ratio of D and G band can be cited to illustrate the graphitic degree and the ID/IG values of S-500, S-600, S-700 and S-800 are 0.96, 1.02, 1.01 and 1.11, respectively. Namely, graphitic degree decreases with the increase of pyrolysis temperature. This phenomenon is common for MOF-derived carbon composites, and the formation of nanocrystalline graphite at higher temperature ought to be the reason for lower graphitic degree [23]. Based on above analysis, S-500, S-600 and S-700 might be (Fe2.5Ti0.5)1.04O4/TiO2/C composites and S-800 may be Fe2TiO4/FeTiO3/C composite. To gather more information about elemental compositions and valence states of the elements, XPS spectra of FTO/C composites have also been supplied. Undoubtably, only peaks of O, Ti, Fe, C elements can be seen in XPS full spectra (Fig. 3c). Considering the existence of Fe-Ti oxides, Fe 2p, Ti 2p and O 1s spectra have been given in Fig. 3d-f. As can be seen in Fe 2p spectra, 2p2/3 peak values of S-500, S-600, S-700 and S-800 appear at 711.08, 711.18, 711.48 and 711.28 eV, and 2p1/2 peaks locate at 724.68, 724.98, 724.98 and 724.98 eV, respectively [24, 25]. Generally, the valence of Fe element would decrease with the increase of pyrolysis temperature for Fe-MOFs, due to carbothermal reduction. However, in this case, the valence of Fe element increases with pyrolysis temperature ranging from 500 oC to 700 oC. The generation of (Fe2.5Ti0.5)1.04O4, instead of Fe0, should promote the abnormal increase of valence for Fe element. We should also mention that compared with (Fe2.5Ti0.5)1.04O4, the formation of Fe2TiO4 and FeTiO3 lead to lower valence of Fe element and thus lower binding energy. As for Ti 2p spectra (Fig. 3e), 2p2/3 peak values of S-500, S-600, S-700 and S-800 occur at 458.48, 458.58, 458.58 and 458.63 eV, corresponding 2p1/2 peaks lie on 464.3, 464.53, 464.58 and 464.53 eV, respectively. Since the valence of Ti element tends to increase with pyrolysis temperature, we may deduce that Ti4+ ions incline to lose electrons in Fe-Ti oxides [26]. Figure 3f shows the O 1s spectra and peaks of S-500, S-600, S-700 and S-800 locate at 530.18, 530.18, 530.33 and 530.33 eV, respectively. Similarly, higher binding energies of samples obtained at higher pyrolysis temperature might prove that O2− ions also tend to lose electrons in Fe-Ti oxides. Therefore, we may infer that Fe ions has a strong preference for capturing electrons in Fe-Ti oxides. Considering that S-800 has a quite different composition, high-resolution XPS spectra of Fe, Ti and O elements were further analyzed by using Lorentzian-Gaussian fitting. In Fig. 3g, the Fe 2p spectrum can be divided into six peaks. The peaks at 710.98 and 724.38 eV are attributed to Fe2+, corresponding to Fe2TiO4 and FeTiO3. The peaks at 712.98, 726.28 and 732.58 eV belong to Fe3+, suggesting possible surface oxidation. Nevertheless, the satellite peak at 719.78 eV may result from Fe0, revealing the fact that not all Fe species are reactive for the generation of Fe-Ti oxides. Moreover, the existence of Fe0 either in oxides particles or carbon frameworks would enhance interfacial polarization. As for Ti 2p spectrum of S-800 (Fig. 3h), the peaks can be well divided into two peaks centered in 458.63 and 464.38 eV, respectively. The distance between two signal peaks is 5.75 eV, which may prove the existence of Ti4+. Figure 3i exhibits the O 1s spectrum of S-800, which shows three peaks at 530.28, 531.23 and 532.36 eV. The first peak is attributed to Ti-O bond and the second peak may originate in surface hydroxyl. The other peak at 532.36 eV illustrates the existence of oxygen vacancy, which would promote dipolar polarization under external electromagnetic field [27].
SEM images of as-synthesized FTO/C composites are depicted in Fig. 4a-d. At lowest pyrolysis temperature of 500 oC, one-dimensional spindle-like structure is well preserved. Compared with Fe-bdc@TiO2, the length and width have decreased to ~ 632 nm and ~ 84 nm, suggesting the decomposition of MOFs. When the pyrolysis temperature increases to 600 oC, the length of spindles continues to decrease to ~ 537 nm. Similarly, S-700 also has spindle-like structure with shorter length of ~ 498 nm. As far as we are concerned, oxides coating layers, including TiO2, SiO2, etc., are always utilized to construct core-shell or york-shell structure [28]. But owing to the solid-state reaction between iron oxides and TiO2, TiO2 shell cannot afford enough structural stability [29]. Consequently, the total length of spindles continues decreasing. It is also worthy pointing out that larger crystals can be seen on the spindle surface, which may be result of the quick random growth of (Fe2.5Ti0.5)1.04O4 particles. When the pyrolysis temperature rises to 800 oC, phase conversion of FTO particles may significantly influence the morphology. As shown in Fig. 4h, S-800 is composed of irregular ellipsoidal particles, whose length is about 151 nm. And depletion of TiO2 shell and generation of new oxide phases should be the reason for structural evolution. Figure 4i shows TEM image of S-700. First, translucent layer, which should be identified as carbon, can be observed around particles. Then, TiO2 shell has already disappeared, and we can conclude that the loosely packed TiO2 particles are able to react with internal iron oxides at a certain temperature. Finally, (Fe2.5Ti0.5)1.04O4 crystals with size of around 30 nm disperse uniformly in the carbon frameworks, creating lots of interfaces. HRTEM image of S-700 is also shown in Fig. 4j. Partially graphitized carbon can be clearly seen among oxide crystals, which is in line with Raman spectrum. One-dimensional porous carbon would play a vital role in building conduction paths both in and between (Fe2.5Ti0.5)1.04O4/TiO2/C particles. As for oxide crystals, the lattice spacing is 0.254 nm, which is consistent with (311) plane of (Fe2.5Ti0.5)1.04O4. Element distribution of S-700 is analyzed, too. According to Fig. 4k, Fe, Ti, O and C elements all disperse uniformly in individual particle. The even distribution of Fe element is inherited from highly ordered molecular structure of Fe-bdc. While even distributions of Ti and O element can be ascribed to both uniform coating of TiO2 and homogeneous reaction between Fe and Ti species. Figure 4l is TEM image of S-800, from which we can see large crystals having thin carbon coating layers. Large crystals correspond well with SEM images, and it can be inferred that high pyrolysis temperature accelerate both phase conversion and crystal growth. According to HRTEM image in Fig. 4m, the lattice spacing is 0.302 nm, which should be attributed to (220) plane of Fe2TiO4. Although phase conversion has taken place, all elements still disperse evenly in individual particle (Fig. 4n). Besides, there are also abundant interfaces in S-800 between oxide particles and carbon framework.
Compared with reported magnetic carbon composites derived from Fe-MOFs, the formation of non-magnetic (Fe2.5Ti0.5)1.04O4, Fe2TiO4 and FeTiO3 would undoubtedly lead to distinct electromagnetic properties. Figure 5a exhibits εr and µr spectra of S-500, and owing to absence of magnetic phases, µ′ and µ″ are close to 1.1 and 0 in whole measuring frequency range, respectively. As for εr, ε′ decreases slowly from 4.41 at 2 GHz to 4.13 at 5.52 GHz and then fluctuates to 4.20 at 14.76 GHz, following by irregular fluctuation. Similarly, the value of ε″ is near 0.30 in frequency range of 2-14.84 GHz, while two adjacent peaks appear between 14.84-18 GHz. Based on εr, S-500 owns low electrical conductivity, although it has high graphitic degree and large aspect ratio. One of the possible reasons could be that residual amorphous TiO2 shells break conductive pathways among spindle-like particles [30]. As for S-600, ε′ goes down from 5.13 at 2 GHz to 3.42 at 18 GHz with three peaks centered in 6.96, 10.6 and 14.64 GHz (Fig. 5b). Meanwhile, ε″ value lies in the range of 0.40–2.02 and three broad peaks at 6.92, 11.88 and 14.84 GHz can be seen. In comparison with S-500, S-600 owns higher ε″, revealing higher electrical conductivity. The continuous loss of amorphous TiO2 shell may promote the formation of conductive pathways and thus improve electrical conductivity. Meanwhile, with continuous generation of (Fe2.5Ti0.5)1.04O4 with low crystallinity and nanocrystalline graphite, defect concentration would also increase, which might promote the development of dielectric relaxation behavior. Besides, µ′ and µ″ of S-600 are also near 1.1 and 0, respectively, showing weak magnetic property. Figure 5c shows εr and µr spectra of S-700. Obviously, ε′ declines steadily from 12.09 at 2 GHz to 4.51 at 18 GHz. While ε″ quickly drops from 7.8 at 2 GHz to 4.78 at 8 GHz and then keeps a value around 4.8. Therefore, S-700 has higher permittivity and electrical conductivity than S-500 and S-600. This can be explained as: (1) amorphous TiO2 shell of spindle-like particles might entirely disappear as shown in Fig. 4f, laying a good foundation for establishing direct contact between individual particles; (2) crystal size of internal (Fe2.5Ti0.5)1.04O4 particles has increased, bringing about content increase of carbon on the surface; (3) graphitic degree of S-700 is relatively high, making surface carbon have enough high conductivity. In addition, S-700 may own weaker dielectric relaxation, which may be associated with higher crystallinity of (Fe2.5Ti0.5)1.04O4 in S-700 than in the others. As for magnetic property, µ′ and µ″ of S-700 are close to 1 and 0, respectively. Figure 5d supplies the electromagnetic parameters of S-800. In detail, ε′ decreases from 12.96 at 2 GHz to 7.22 at 18 GHz and ε″ goes down from 7.51 at 2 GHz to 2.92 at 18 GHz. Moreover, dielectric relaxation occurs in frequency range of 6–8 and 10–12 GHz, respectively. On one hand, S-800 has lower electrical conductivity than S-700, especially at high frequency. According to SEM images, S-800 is composed of ellipsoidal particles with shorter length, which may increase the difficulty for establishing direct contact among individual particles. On the other hand, S-800 owns stronger dielectric relaxation, which may correlate with the multi-phases. After the conversion from (Fe2.5Ti0.5)1.04O4 to Fe2TiO4 and FeTiO3, more interfaces and crystal defects may be generated, which would induce strong polarization. Similar with S-700, S-800 is also nonmagnetic material, which has low µr close to air. According to above analysis, more attention should be paid to electrical conductivity, which leads to great differences in complex permittivity. Based on characterization results and theoretical research, Fig. 5e further summarizes the relations between microstructure and electrical conductivity. First, conventional MOF-derived structure of carbon composites can be described as metal oxide or metal nanoparticles distribute uniformly in porous carbon matrix. Obviously, conductive carbon is the prerequisite for constructing electrically conductive path in induvial particle and between particles. Under the influence of external electrical field, free electrons would move directionally in carbon matrix [31]. When the internal particles are highly resistive, moving electrons in carbon would be scattered by uniformly dispersed particles, leading to significantly decreased electron mean free path. This may be one of the important reasons for low conductivity of S-500 and S-600. When uniformly dispersed small particles become large particles, which distribute in the center of porous carbon rod, moving electrons would travel quickly in the continuous conductive surface carbon layer, resulting in high conductivity of S-700 and S-800. Second, one-dimensional conductive carbon frameworks can connect with each other easily in paraffin matrix, owing to low percolation threshold and absence of insulation layer. This can explain why S-700 owns highest conductivity. On the contrary, the collapse of one-dimensional structure for S-800 and the existence of TiO2 insulation layer for S-500 and S-600 should be blamed for low electrical conductivity.
Dielectric loss tangents (tan δE) of FTO/C composites have been displayed in Fig. 6a. We can see that S-700 has highest tan δE in whole measuring frequency range, revealing strongest dielectric loss. As discussed above, conduction loss generated in carbon frameworks plays a key role in dielectric loss of S-700. Besides, interfacial polarization arising in interface region between (Fe2.5Ti0.5)1.04O4 and carbon is also involved in dielectric loss [32]. Similarly, S-800 owns the second highest tan δE in almost whole testing frequency range, which can be understood by high electrical conductivity of S-800. It is also worth mentioning that dielectric relaxation enhances dielectric loss at frequency near 11 GHz. We should also point out that S-600 has even higher tan δE than S-800 in frequency range of 11.76–12.72 GHz and strong dielectric loss induced by dipolar polarization should be the main reason [33]. Besides, S-500 shows lowest tan δE, which should be related to the presence of amorphous TiO2 shell. To sum up, higher pyrolysis temperature contributes to the loss of amorphous TiO2 shell and the exposure of conductive carbon, which are crucial in enhancing dielectric loss. Meanwhile, pyrolysis temperature also poses significant effect on dipolar and interfacial polarization loss. Magnetic loss tangents (tan δM) are also displayed in Fig. 6b. For one thing, S-600, S-700 and S-800 all show negative tan δM values at high frequency, and this phenomenon means magnetic energy is radiated out and transformed into electric energy [33]. For another, S-500 and S-600 own slightly higher tan δM than S-700 and S-800. It can be inferred that the reaction between Fe-based species and TiO2 is still incomplete in S-500 and S-600 and the possible existence of magnetic phase (for example, Fe) may induce slightly stronger magnetic loss. But in general, FTO/C composites all own much weaker magnetic loss than dielectric loss and should be regarded as dielectric MAMs. To gain more insights into dielectric loss mechanisms of S-700 and S-800, Cole-Cole plots have been depicted as Fig. 6c, d. According to Debye theory,
$${\epsilon }^{{\prime }}={\epsilon }_{\infty }+\frac{{\epsilon }_{s}-{\epsilon }_{\infty }}{1+{\left(2\pi f\right)}^{2}{\tau }^{2}}$$
3
$$\epsilon {\prime }{\prime }={\epsilon }_{\infty }+\frac{2\pi f\tau ({\epsilon }_{s}-{\epsilon }_{\infty })}{1+{\left(2\pi f\right)}^{2}{\tau }^{2}}$$
4
where ε∞, εs, τ are optical dielectric constant, static dielectric constant and relaxation time, respectively. Based on above equations, one can deduce that
$${({\epsilon }^{{\prime }}-\frac{{\epsilon }_{s}-{\epsilon }_{\infty }}{2})}^{2}+{\left({\epsilon }^{{\prime }{\prime }}\right)}^{2}={\left(\frac{{\epsilon }_{s}-{\epsilon }_{\infty }}{2}\right)}^{2}$$
5
Therefore, the ε′-ε″ curve, also known as Cole-Cole plot, should be a semicircle if only one type of Debye relaxation has occurred [34]. As can be seen in Fig. 6c, Cole-Cole plot of S-700 is made up of four incomplete semicircles and a straight tail. So, four types of Debye-like polarization have taken place in S-700. In terms of microstructure, lot of crystals disperse uniformly in porous carbon rods, and thus create abundant interfaces, in which diverse types of polarization may be induced. While for S-800 (Fig. 6d), three incomplete semicircles and a straight tail can be viewed, revealing only three types of polarization processes. Although S-800 shows fewer types of polarizations, it also shows strong dielectric relaxation. In our opinion, carbon in S-800 has lowest graphitic degree and the thermal treatment time may be too short to repair the defects in newly generated Fe2TiO4 and FeTiO3. Moreover, phase conversion of Fe-Ti oxides also brings about more interfaces. Hence, strong defect-induced polarization and interfacial polarization would occur.
Due to the absence of magnetic loss, impedance matching conditions should be paid more attention to, when FTO/C composites are used as MAMs. Generally, when impedance matching ratio (Z = Zin/Z0) equals 1, perfect impedance matching would be realized. And when Z value lie in the range of 0.52–1.93, effective impedance matching can be achieved. As shown in Fig. 7a, fan-shaped effective impedance matching area has been separated into two isolated parts. Therefore, when the thickness is fixed (> 1.95 mm for S-500), continuous impedance matching cannot be obtained over a wide frequency range, which is against critical requirement of broadband microwave absorption. Similar phenomenon has occurred for S-600, as revealed by Fig. 7b. Interestingly, good impedance matching can be always realized around 11.6 GHz. This may be explained as the increase of ε″ and the decrease of ε′ are helpful to the optimization of impedance matching. In other words, dielectric relaxation may make material receive and consume more incident wave in a certain situation [35]. As shown in Fig. 7c, when the thickness ranges from 1.45 to 2.2 mm, continuous impedance matching can be got in whole Ku band, and when the thickness is larger than 2.05 mm and smaller than 2.2 mm, satisfied results would be acquired in whole Ku and X bands. Similarly, S-800 also shows excellent impedance matching (Fig. 7d). For example, when the thickness lies in the range of 1.25-2 mm, effective impedance matching can be achieved in Ku band. Moreover, when the thickness is only 2 mm, suitable impedance can be realized in both Ku and X bands. Considering FTO/C composites are dielectric materials, S-700 and S-800 may have proper εr, owing to suitable electrical conductivity and various polarization processes. In contrast, S-500 and S-600 may own too low εr, which should be ascribed to low electrical conductivity.
Reflection loss performance of as-prepared FTO/C composites has been supplied as Fig. 8a-d. As shown in Fig. 8a, RL peak value of -8.06 dB can be reached at 15.32 GHz with a thickness of 2.8 mm. Weakest attenuation ability should be the main reason for poor microwave absorption property of S-500. Similarly, S-600 is also not capable of absorbing incident microwave in a wide frequency range, due to the lack of strong dielectric loss (Fig. 8b). Howbeit, EAB of 2.08 GHz (10.72–12.8 GHz) can be achieved at 3.4 mm with RL peak value of -16.27 dB. As discussed above, dielectric relaxation not only enhance the attenuation ability, but also improve impedance matching, thus guaranteeing effective microwave absorption. Figure 8c exhibits the microwave absorption performance of S-700. Thanks to strongest dielectric loss and good impedance matching, a broad EAB of 6.84 GHz can be realized at thickness of 2.2 mm. It should be noted that Ku band can be entirely covered within a thickness range of 2.1–2.2 mm as shown in Fig. 8e. S-800 also shows good microwave absorption performance (Fig. 8d). A broad EAB of 4.76 GHz can be obtained at 1.8 mm with smallest RL value of -25.29 dB. Although S-800 also has good impedance matching behavior in Ku band, inadequate attenuation ability may be responsible for narrower absorption bandwidth in comparison with S-700. Howbeit, S-800 owns better microwave absorption performance in X band. In detail, effective absorption can be got from 7.88 GHz to 12.04 GHz at thickness of 2.8 mm, covering whole X band (Fig. 8f). Since dielectric loss of S-800 is weaker than that of S-700, wider absorption bandwidth in X band may be ascribed to proper impedance. Different from absorption bandwidth, interference cancelation between reflected wave from the surface of absorber and the surface of metal back decides the RL peak frequency, which can be explained by a quarter-wavelength matching model:
\({d}_{m}=\frac{n\lambda }{4}=\frac{nc}{4{f}_{m}\sqrt{\left|{\mu }_{r}\right|\left|{\epsilon }_{r}\right|}}\) (n = 1, 3, 5, …) (6)
Here, dm and fm represent the matching thickness and interference frequency [36]. To further clarify the relations among RL, impedance matching ratio and interference cancelation, 2.05 and 2.2 mm are selected for S-700. As shown in Fig. 8g, 2.2 mm corresponds to largest EAB, and 2.05 mm is the smallest thickness when effective impedance matching can be gained in whole Ku and X bands. On one hand, good impedance matching in a wide frequency range is necessary for the extension of effective absorption bandwidth, especially when Z approaches 1. But effective absorption is also decided by adequate attenuation ability. That is why effective absorption cannot be realized in X band at above selected thickness. On the other hand, Z values at RL peak frequencies are not 1 in this case. This gap may be also related with different attenuation ability. Similarly, RL peak frequencies are not perfectly in line with interference frequencies. First, interference cancellation plays a vital role in generation RL peak. Then, compared with theoretical predication, at the same matching thickness, RL peak appears at higher frequency. This can also be explained as stronger dielectric loss at high frequency promotes the increase of RL peak frequency [37]. With special emphasis on X band, 2 mm and 2.8 mm have been chosen for S-800. Interestingly, Z values at RL peak frequencies are remarkably close to 1 as exhibited in Fig. 8h. Namely, perfect impedance matching poses significant effect on generating RL peak. However, at the same matching thickness, RL peak frequencies are also a bit higher than interference frequencies, which should also be in association with attenuation ability. Overall, S-700 and S-800 have excellent microwave absorption performance and are comparable with reported MAMs. Figure 8i summarizes RL peak values of similar MAMs. Minimum RL value of -42.38 dB can be reached at thickness of 2.4 mm for S-800, which can meet most practical situations [24, 25, 29, 38–42]. Figure 8j lists EAB of reported MAMs. Obviously, S-700 has a wide EAB of 6.84 GHz, which is larger than most similar MAMs [24, 25, 29, 38–42]. To sum up, as-prepared FTO/C composites can be served as excellent MAMs, especially in Ku and X bands.
According to above analysis, microwave absorption mechanisms have been displayed in Fig. 9. First, impedance matching, which is the precondition of effective microwave absorption, should be emphasized. Owing to suitable ε′ and ε″, S-700 and S-800 own appropriate Zin in a wide frequency range, laying a good foundation for broadband microwave absorption. Then, interference cancellation occurring at specific frequency plays a decisive part in generating reflection loss peak. Due to high εr, both S-700 and S-800 have thin matching thickness, especially at high frequency. Third, under the influence of external electric field, migrating and hopping electrons in graphitic carbon framework would produce heat. In this case, grain size increase and concentrated distribution of internal FTO crystals, as well as one-dimensional morphology of individual FTO/C particles, promote the formation of large-scale electrically conductive networks, thus enhancing the conduction loss. Fourth, dipoles would be produced either in carbon or in FTO crystals, which would constantly turn their direction under the influence of alternating electric field. As a result, incident microwave energy would be consumed by dipolar polarization. Besides, due to different dielectric properties, space charge carriers would accumulate in interfaces among FTO crystals and carbon shells. Namely, interfacial polarization may be also involved in the microwave attenuation.