Although many materials have been studied for the purpose of microwave absorption, SiO2 has never been reported as a good candidate. In this study, we present for the first time that doped, microwave conductive SiO2 nanoparticles can possess an excellent microwave absorbing performance. A large microwave reflection loss (RL) of −55.09 dB can be obtained. The large microwave absorption originates mainly from electrical relaxation rather than the magnetic relaxation of the incoming microwave field. The electrical relaxation is attributed to a large electrical conductivity that is enabled by the incorporation of heterogeneous (N, C and Cl) atoms. The removal of the magnetic susceptibility only results in a negligible influence of the microwave absorption. In contrast, the removal of the heterogeneous atoms leads to a large decrease in the electrical conductivity and microwave absorption performance. Meanwhile, the microwave absorption characteristics can be largely adjusted with a change of the thickness, which provides large flexibility for various microwave absorption applications.
The utility of microwave absorption is significant in many applications, such as wireless communications and anti-radar detection of aircraft1,2,3. For example, many materials have been developed to reduce the radar signature of aircraft, ships and tanks in military fields. Material examples include carboneous materials such as graphite4, graphene5, carbon nanotubes (CNTs)6 and carbon fibers7, conducting polymers8, oxides such as Fe2O39, Fe3O410, MnO211, ZnO12, BaFe12O1913, BaTiO313, SrFe12O1914, and carbides such as SiC15 and SiCN16. Traditional mechanisms that are commonly believed to be responsible include dipole rotation and magnetic domain resonance due to the dielectric and magnetic losses inside the materials. For example, the microwave absorption of Ni-coated CNT/epoxy composites results from dielectric and magnetic losses, with the use of Ag nanowires shown to enhance the performance due to dielectric loss3. Conducting polymers such as polypyrrole, polyaniline and poly(3-octylthiophene) have been demonstrated to show good microwave absorption8. Suitable inclusion of Fe2O3 nanoparticles in polyaniline enhanced the microwave absorption properties due to simultaneous adjustment of dielectric loss and magnetic loss9. Hollow urchin-like MnO2 nanostructures with tetragonal nanorod clusters showed better performance due to proper electromagnetic impedance matching11. BaTiO3/polyaniline and BaFe12O19/polyaniline composites showed good compatible dielectric and magnetic properties for broadband microwave absorbing properties13. Recently, microwave absorption has been reported to be enhanced by defective or disorder structures such as oxygen vacancies, low crystallinity and use of a core/shell structure1,2,17,18,19. The dielectric properties of TiO21,2,17,18, ZnO2 and BaTiO319 nanoparticles can be modified through the perturbation of their crystalline structure by hydrogenation treatment to enhance their microwave absorption performance. Gao and colleagues20 have also found that the microwave dielectric properties of TiO2 nanoparticles can be largely altered by introducing partial crystalline phases in such TiO2 nanoparticles. Although these findings have provided a large pool of materials for microwave absorption applications, each material may have some advantages and disadvantages; therefore, it is highly desirable to discover new materials for microwave absorption. On the other hand, microwave absorption is also very useful for sample preparation with microwave irradiation in that rapid synthesis is enabled when compared to use of a traditional heating treatment21,22.
Up to now, pure SiO2 has not yet been reported to show a good microwave absorption performance, although an enhanced microwave performance for the composites of SiO2/carbon23, SiO2/carbonyl iron24 and SiO2/iron25 was obtained due to the combination of a proper electromagnetic impedance match23,24,25. This is understandable because in pure SiO2 there is no good source for traditional microwave absorption mechanisms including dipole rotation and magnetic domain resonance due to dielectric and magnetic losses. The lack of origin for the creation of dipoles in pure SiO2 nanoparticles is attributed to its symmetric structure where the dipoles are canceled in each tetrahedral unit of the SiO2 lattice. Meanwhile, it is also unlikely that there will be magnetic domains to echo with the microwave electromagnetic field due to the lack of charge spin centers in pure SiO2. Therefore, it is reasonable that pure SiO2 will not be able to possess good mechanisms for efficient microwave absorption. To enable effective microwave absorption, we intentionally create dipoles in SiO2 nanoparticles by introducing heterogeneous atoms such as C, N and Cl elements on the O sites inside the SiO4 tetrahedra or simply linked onto the surface of the SiO2 nanoparticles. Rotations of these induced dipoles can, therefore, echo with the incident microwave field to possess microwave absorption. Effective electrical loss can be achieved with such heterogeneous atoms by having higher electrical conductivity. Here, we report that the doped, microwave conductive SiO2 nanoparticles can show excellent microwave absorption performance. In contrast, only poor microwave absorption performance is observed for the SiO2 nanoparticles without the incorporation of the heterogeneous atoms, which lack dipole rotations and good electrical conductivity.
Materials and methods
Tetraethyl orthosilicate (TEOS), N,N’-dimethylformamide (DMF) and hydrazine monohydrochloride were purchased from Sigma-Aldrich and used as received. Typically, TEOS (2.2 mL) was slowly added to DMF (20.0 mL) under ambient conditions by vigorously stirring, producing a completely transparent solution. A desired amount of hydrazine monohydrochloride (2.74 g) was added and stirred for 5–10 min, which was then transferred into a Teflon-lined, stainless autoclave at a different temperature of 160 °C for 12 h, producing a light-yellow slurry, which was washed with anhydrous ethanol and dried at 100 °C. A light-yellow product was collected. For comparison, pure white SiO2 nanoparticles were also prepared followed by calcination of the light-yellow sample at 600 °C in air for 2 h.
The crystal structure was examined using a Rigaku Miniflex X-ray diffractometer (XRD) with a Cu Kα (λ = 0.15418 nm) radiation source. The morphology and crystallinity were probed with transmission electron microscopy (TEM) and high-resolution TEM (HRTEM) on an FEI Tecnai F20 STEM under an electron accelerating voltage of 200 kV. The surface chemical properties were studied with a Kratos Axis 165 X-ray photoelectron spectrometer (XPS) equipped with an Al/Mg dual-anode X-ray source. All the XPS spectra were calibrated with the C 1s peak from the carbon tape to 284.6 eV. The energy-dispersive X-ray (EDX) spectrum was taken using a Tescan Vega 3 LMU scanning electron microscope equipped with a Bruker Quantax 6|10 EDX system. The Fourier transform infrared (FT-IR) spectra were collected using a Thermo-Nicolet iS10 FT-IR spectrometer equipped with an attenuated total reflectance unit. The complex permittivity and permeability were measured in the frequency range of 1.0–18.0 GHz using an HP8722ES network analyzer with ring-shape samples containing 60 wt% SiO2 nanoparticles dispersed in paraffin wax. The ring had a thickness of 2.0 mm, an inner diameter of 3.0 mm and an outer diameter of 7.0 mm.
Results and discussion
The XRD pattern for the doped SiO2 nanoparticles matched well with the standard card of SiO2 (PDF#00-038-0360), as shown in Fig. 1b. The broadness of the broad peak at approximately 26.2° indicated that the nanoparticle size was small. The weak intensity suggested a poor crystallinity. From the TEM image in Fig. 1c, clearly, the nanoparticles were 4–8 nm in diameter (Fig. 1d). The HRTEM image in Fig. 1e exhibited clear lattice fringes with a plane distance of 0.245 nm corresponding to the (200) plane of monoclinic Moganite SiO2. Meanwhile, some amorphous regions were also observed between the crystalline domains, suggesting the likely existence of an amorphous phase (as pointed out by the dashed lines). The small size and amorphous phase matched well with the broad diffraction peak from the XRD pattern. This coexistence of amorphous and crystalline domains in the doped SiO2 nanoparticles was similar to that reported for TiO2 nanoparticles showing enhanced microwave absorption;1,2,20 this is because the amorphous phase may create some interfacial dipole rotations along the interface with the crystalline phase and induce active microwave absorption1,2,20. The existence of Si, O, C, N and Cl elements were confirmed using results from XPS and EDX, as shown in Figure S1-S7. The surface of the doped SiO2 nanoparticles was likely linked with the –NH2 groups from the hydrazine hydrochloride and some adsorbed water and HCl molecules (Figure S8). The Si 2p spectrum (Figure S2) showed one peak with a binding energy centered approximately 103.2 eV, corresponding to the Si 2p3/2 in the SiO226,27. In the O 1s XPS spectrum shown in Figure S3, one peak was found to be centered approximately 532.6 eV, consistent with the binding energy of the O 1s signal from SiO228,29. The C 1s XPS spectrum (Figure S4) showed one minor peak centered approximately 284.6 eV, and one major peak approximately 286.4 eV. The former was likely from the adsorbed carbon during the XPS measurement, with the latter likely from the alkyl groups from the TEOS on the surface of the SiO2 nanoparticles30,31. The N 1s spectrum in Figure S5 showed two peaks with a major contribution approximately 399.9 eV and a minor shoulder approximately 401.9 eV, likely from the N2H4 moiety and the conjugated NH2 moiety attached to the HCl group from the hydrazine hydrochloride on the surface of the SiO2 nanoparticles, respectively32,33. The Cl 2p XPS spectrum (Figure S6) displayed one higher peak approximately 197.6 eV and one lower peak near 199.3 eV, likely from the HCl coupled to the N2H4 moiety on the surface and the HCl coupled to the SiO2 surface, respectively34.
The variations of the microwave reflection loss (RL) with frequency (f) and thickness (d) were clearly displayed in a three-dimensional (3D) plot and two-dimensional (2D) contour, as shown in Fig. 2a, b. The 2D contour plot showed the projection of the 3D graph for the change of RL (indicated by the color ruler) on the frequency f and thickness d plane. The different colors indicated where (f and d) and which level of RL was achieved. Some representative RL curves were shown in Fig. 2c for varying d from 1.0 to 20.0 mm. The maximum microwave absorption frequency (fmax) was tunable with d. As d increased, fpeak shifted from higher frequency to lower frequency, as clearly seen in Fig. 2d. The relationship between fpeak and d was fitted very well with the formula f/GHz = 39.0/(d/mm)1.17 or c/4fpeakε’0.5 = λ/4ε’0.5, where c was the speed of light. Apparently, fpeak decreased reversibly with d. As d became bigger, fpeak decreased. The change in RLpeak with d (Fig. 2e) could be divided into two stages: a quick increase from −18.45 to −55.09 dB when d was changed from 2.0 to 4.2 mm, and a decay from −55.09 to −9.97 dB as d grew from 4.2 to 20.0 mm. The largest RLmax value (−55.09 dB) was observed when d was 4.2 mm. The change in Δf10 with d was shown in Fig. 2f and fitted very well with Δf10 /GHz = 12.4/(d/mm)0.68 – 1.65. As d became bigger, Δf10 almost decreased monotonically. This indicated that a thicker coating actually shielded a narrower region of microwave reflection frequency. It should be noted that the application of a thicker coating is very useful for the protection of important stationary objects on the ground from radar detection where a thin coating may not be able to shield such objects at a specific frequency, which is very important but frequently overlooked.
The microwave absorption properties are dependent on the dielectric and magnetic properties:
where RL(dB) is the reflection loss in dB, Zin is the input impedance of the absorber, Z0 is the impedance of free space, μr is the relative complex permeability, εr is the relative complex permittivity, f is the frequency of the electromagnetic wave, d is the thickness of the absorber and c is the velocity of light18. As shown in Fig. 3a, when f increased from 1.0 to 18.0 GHz, ε’ gradually decreased from 9.89 to 5.35, ε” decreased slowly from 7.63 to 2.06 and tgδε fell slowly from 0.77 to 0.39. Figure 3b showed that μ’ decreased from 1.05 to 0.99, μ” dropped from 0.06 to 0.03 and tgδμ changed from 0.06 to 0.03 when f increased from 1.0 to 18.0 GHz. These results suggested that the doped SiO2 nanoparticles had a smaller stored electrical and magnetic energy as the frequency of the incident electromagnetic field increased, indicating that some of the echoes of the electric field or dipoles to the oscillating field lagged behind and seemed consumed as the frequency increased.
The electrical conductivity (σ) shown in Fig. 3c was calculated with σ (S m−1) = 2πfε0ε”, where ε0 was the free space permittivity (8.854 × 10−12 F m−1), f was the frequency (Hz) and ε” was the imaginary component of the permittivity35. The σ increased almost monotonically from 0.42 to 2.06 S m−1 as f increased from 1.0 to 18.0 GHz. The large conductivity was suggested to be possibly related to the existence of the heterogeneous atoms (C, N, Cl) on the surface of the SiO2 nanoparticles; for example, partial oxygen atoms were replaced with N atoms on the surface of the particles. The skin depth (δ) of the microwave irradiation in Fig. 3d was calculated with (δ/m) = (πfμ0μrσ)−1/2, where μ0 was the permeability of free space (4π × 10−7 H m−1), μr was the relative permeability and σ was the electrical conductivity (S m−1)33. The δ decreased from 23.84 to 2.63 mm as f increased from 1.0 to 18.0 GHz.
To reveal the contribution of the permeability or the magnetic properties of the SiO2 nanoparticles to the microwave absorption performance, we compared the microwave absorption results with and without the contribution of the magnetic components by assuming, for the latter case, a magnetic susceptibility parameter (χm) equal to zero, where μr = μ/μ0 = (1 + χm)μ0. The evolution of the RL curves in Fig. 4a when d was changed from 1.0 to 20.0 mm indicated that as d increased, fpeak shifted to lower values. As shown in Fig. 4b, fpeak decreased as d increased, with their relationships overlapping almost completely for nonzero and zero χm (see also Figure S9). RLpeak rapidly increased and then decayed with d, following the same trend for nonzero χm (Fig. 4c and Figure S10). Meanwhile, it was noticeable that the maximum RLpeak was smaller with a zero χm as d changed. Figure 4d compared the Δf10–d relationships for nonzero and zero χm. The Δf10–d trends overlapped very well (also see Figure S11) despite Δf10 being smaller at smaller d values and larger at larger d values for zero χm. This indicated that the influence of the none-zero χm for the SiO2 nanoparticles was mainly reflected in the change of the Δf10 values. Overall, the none-zero χm increased the RLpeak and the Δf10 values possibly achieved at certain d values despite the overall impact being small. The small contribution of the magnetic property on the microwave absorption was related to the small μ” and tgδμ values. The overall influence of the none-zero χm on the microwave absorption was clearly shown in Figure S12 as well.
Therefore, the large microwave absorption performance of the doped, conductive SiO2 nanoparticles was most likely related to their dielectric properties, or the rotations of dipoles in the material as indicated by the large electrical conductivity in the microwave frequency range, as shown in Fig. 3c. In pure SiO2 nanoparticles, no obvious origin was found for the creation of dipoles due to its symmetric structure where the dipoles likely canceled out each other in each tetrahedral unit of the SiO2. However, in the doped SiO2 nanoparticles, there were possible sources for the existence of dipoles due to the introduction of C, N and Cl elements as evidenced by the XPS results. Those atoms apparently broke down the symmetrical environment of the SiO2 lattice, and created dipoles on the surface, causing the variation of the dielectric property across the nanoparticles and the increased conductivity. Under microwave irradiation, those dipoles might rotate and echo with the electromagnetic field to produce resonance, causing reflection loss, as schematically shown in Figure S13. Therefore, a large RL was observed in the case of a good match between the dielectric/magnetic properties with the incident microwave field.
To verify this conclusion, we measured the microwave absorption performance along with the dielectric and magnetic properties of those doped SiO2 nanoparticles after removal of the heterogeneous atoms by calcination at 600 °C for 2 h in air. The removal of those atoms was confirmed using XPS results (Figure S14-17) where no N and Cl atoms were observed, with the remaining C due to atmospheric deposition. Meanwhile, calcination at high temperature normally led to high crystallization and removal of the amorphous phase in the material. Compared to the doped SiO2 nanoparticles, the calcinated, pure SiO2 nanoparticles showed much smaller ε’, ε” and tgδε values (Figure S18), which matched well with the values in the literature36, but showed similar μ’, much larger μ” and tgδμ values (Figure S19). As a result, these calcinated, pure SiO2 nanoparticles only showed very small σ values (Fig. 5a) and large δ (Fig. 5b) values in the microwave region. The small σ values indicated that these SiO2 nanoparticles were barely electrically conductive, with the large δ values suggesting that the microwave field was not efficiently decayed. Their poor microwave absorption performance was observed from the 2D contour plot of the RL over f and d in Fig. 5c and the RL plots in Fig. 5d, where the RL values were found to be less than −10 dB in most of the d–f regions. The poor microwave absorption performance of the calcinated SiO2 nanoparticles further confirmed that the good microwave absorption performance of the doped SiO2 nanoparticles was due to electrical relaxation, as the calcinated SiO2 nanoparticles had much smaller electrical relaxation but much large magnetic relaxation. Meanwhile, the fact that the associated C, N and Cl atoms were removed from the SiO2 nanoparticles after calcination also hinted that those heterogeneous atoms were likely linked onto the surface instead of being located in the bulk of the SiO2 nanoparticles. Table S1 listed the microwave absorption performance of various materials that have been studied. As seen, doped, microwave conductive SiO2 showed an impressive microwave absorption performance. Therefore, such materials are a promising candidate for microwave absorption. The importance of electrical relaxation on the microwave absorption is consistent with the conclusion made in the studies by Mao and colleagues37,38,39,40,41,42,43. Furthermore, such materials can be shown to be very useful in self-powered electromagnetic energy conversion and microwave attenuation, as demonstrated by Mao and colleagues40,41, which we plan to build in our future work.
In summary, in this study we have shown that doped, microwave conductive SiO2 nanoparticles possess an exciting microwave absorbing performance, benefiting from the good electrical conductivity that is induced by the incorporation of heterogeneous (N, C and Cl) atoms on the surface of the SiO2 nanoparticles. A large RL value of −55.09 dB can be obtained. The microwave absorption is mainly related to the dielectric loss resulting from the good electrical conductivity, while the small contribution of the magnetic property may be directly related to the small μ” and tgδμ. In contrast, calcinated, pure SiO2 nanoparticles show a poor electrical conductivity and microwave absorption performance even with a larger magnetic response. In addition, microwave absorption characteristics such as fpeak, RLpeak and Δf10 can be largely adjusted with the thickness of the doped, conductive SiO2 nanoparticles, which provides large flexibility for their microwave application towards various purposes.
X.C. appreciates the support from the U.S. National Science Foundation (DMR-1609061) and the College of Arts and Sciences, University of Missouri–Kansas City. L.L. acknowledges the support from the National Science Fund for Distinguished Young Scholars of China (No. 61525404). M.Z. acknowledges the support from the National Natural Science Foundation of China (Grant No. 51372080). X.T. is thankful for the support from the National Natural Science Foundation of China (U1765105). F.H. appreciates the support from the National Key Research and Development Program of China (2016YFB0901600).