Research Article www.acsami.org
Enhanced Microwave Absorption Performance of Coated Carbon Nanotubes by Optimizing the Fe3O4 Nanocoating Structure Na Li,† Gui-Wen Huang,† Yuan-Qing Li,*,‡ Hong-Mei Xiao,† Qing-Ping Feng,† Ning Hu,‡ and Shao-Yun Fu*,†,‡ †
Technical Institute of Physics and Chemistry, Chinese Academy of Sciences, Beijing 100190, China College of Aerospace Engineering, Chongqing University, Chongqing 400044, China
‡
S Supporting Information *
ABSTRACT: It is well accepted that the microwave absorption performance (MAP) of carbon nanotubes (CNTs) can be enhanced via coating magnetic nanoparticles on their surfaces. However, it is still unclear if the magnetic coating structure has a significant influence on the microwave absorption behavior. In this work, nano-Fe3O4 compact-coated CNTs (FCCs) and Fe3O4 loose-coated CNTs (FLCs) are prepared using a simple solvothermal method. The MAP of the Fe3O4-coated CNTs is shown to be adjustable via controlling the Fe3O4 nanocoating structure. The results reveal that the overall MAP of coated CNTs strongly depends on the magnetic coating structure. In addition, the FCCs show a much better MAP than the FLCs. It is shown that the microwave absorption difference between the FLCs and FCCs is due to the disparate complementarities between the dielectric loss and the magnetic loss, which are related to the coverage density of Fe3O4 nanoparticles on the surfaces of CNTs. For FCCs, the mass ratio of CNTs to Fe3+ is then optimized to maximize the effective complementarities between the dielectric loss and the magnetic loss. Finally, a comparison is made with the literature on Fe3O4-carbon-based composites. The FCCs at the optimized CNT to Fe3+ ratio in the present work show the most effective specific RLmin (28.7 dB·mm−1) and the widest effective bandwidth (RL < −10 dB) (8.3 GHz). The excellent MAP of the as-prepared FCC sample is demonstrated to result from the consequent dielectric relaxation process and the improved magnetic loss. Consequently, the structure−property relationship revealed is significant for the design and preparation of CNT-based materials with effective microwave absorption. KEYWORDS: carbon nanotube, iron oxide, microwave absorption, coating structure, reflection loss, impedance matching
1. INTRODUCTION The development of high-performance microwave absorption materials has become a recent focus in resolving the electromagnetic interference (EMI) pollution arising from the fast-growing electronic devices industry.1−3 Ideal microwave absorption materials are required to have low density and low thickness with strong absorption over a broad frequency range. Recently, an ultralight and highly compressible carbon-based material with excellent microwave absorption performance (MAP) and very wide effective frequency bandwidth has been reported.4 Carbon nanotubes (CNTs), which are lightweight and possess excellent electrical conductivity and mechanical properties, are also highly attractive in the field of microwave absorption.5−7 It is known that the microwave loss mechanism includes dielectric loss and magnetic loss, and effective complementarities between these are necessary to achieve excellent MAP. However, microwave absorbents with a single composition and structure are only based on one sort of loss mechanism.2,8,9 For CNTs, the loss mechanism is mainly based on dielectric loss that results from polarization, so their microwave absorption properties are insufficient to broaden their applications.10 To enhance the magnetic loss, it is feasible © 2016 American Chemical Society
to combine magnetic particles with CNTs. Through the introduction of magnetic metals11−14 or magnetic metal oxides,15−19 the electromagnetic shielding and attenuation performance of CNTs can be significantly improved, which makes them ideal candidates as microwave absorption materials.20 CNT/Fe, CNT/Co, and CNT/Ni nanocomposite powders have been prepared by simple chemical methods and show good microwave absorption properties due to the combination of permeability and permittivity resulting from the magnetic nanoparticles and CNTs.12,21 Due to its low toxicity, high compatibility, and strong spin polarization at room temperature, ferroferric oxide (Fe3O4) is highly attractive for enhancing the magnetic loss and electromagnetic attenuation of CNTs or carbon-based materials.2,15,21−23 Highly efficient microwave attenuation has been observed in Fe3O4-decorated CNTs.15 The interface introduced by Fe3O4 generates resonance in complex permittivity and permeability as well as enhanced magnetic loss, which results in the Received: October 28, 2016 Accepted: December 27, 2016 Published: December 27, 2016 2973
DOI: 10.1021/acsami.6b13142 ACS Appl. Mater. Interfaces 2017, 9, 2973−2983
Research Article
ACS Applied Materials & Interfaces
Figure 1. TEM, schematic diagram (top left corners of the insets), and HRTEM (top right corners of the insets) images of (a) FCCs and (b) FLCs. Microwave absorption characteristics of (c) FCCs and (d) FLCs at 2−18 GHz with different thicknesses (30 wt % in paraffin composite). Frequency dependence of complex permittivity and permeability, dielectric tangent loss, and magnetic tangent loss of (e, g) FCCs and (f, h) FLCs.
enhanced MAP and widened effective absorption bandwidth.15 Therefore, Fe3O4/CNT composites are considered to be a promising candidate for microwave absorption. Several factors, such as morphology, geometry, and microstructure, are vital in determining the microwave absorption behavior of absorbers.24,25 The effects of the CNT structure, including CNT aspect ratio and wall integrity, on the EMI shielding effectiveness (SE) have been studied, and the results show that a higher aspect ratio and better surface integrity of the CNTs give a higher EMI.26 On the other hand, microwave absorption property investigations of single-crystal Fe3O4, γFe2O3, and Fe micropine dendrites demonstrate that their crystallization structures and coercivity values cause different microwave absorption effectivenesses.27 Furthermore, Fe3O4
nanoparticles with different diameters28−30 and Fe3O4 nanorings with different long axis values31 also lead to varying microwave absorption results. The efficient low-frequency MAP of certain Fe3O4-C nanorings has been demonstrated to be attributable to their core−shell structure.8 For the Fe3O4-coated CNTs, the coating structure of Fe3O4 is significant in determining the interfacial impedance matching and complementarities between the dielectric loss and the magnetic loss, which are key factors to the microwave absorption properties.2,32 However, the effect of magnetic coating structure on the microwave absorption effectiveness of Fe3O4-coated CNTs is not yet clear. To further enhance their microwave absorption properties, it is critical to study the effect of the coating 2974
DOI: 10.1021/acsami.6b13142 ACS Appl. Mater. Interfaces 2017, 9, 2973−2983
Research Article
ACS Applied Materials & Interfaces
spectroscopy (EDS) analysis was carried out on a HITACHI S-4800 electron microscope (Japan) operated at 15 kV acceleration voltage. The phase of the products was characterized by an X-ray diffractometer (XRD) with Cu Kα radiation (λ = 0.154 nm) between 2θ = 10° and 2θ = 90°. Thermogravimetry (TG) (TA-5000 apparatus, USA) was conducted on Fe3O4-coated CNTs in platinum crucibles under air atmosphere at a heating rate of 10 K/min. The complex permeability and permittivity of Fe3O4-coated CNTs were measured by dispersing the samples in paraffin wax with a mass ratio of 30 wt % and then pressing the mixture into a toroidal shape (inner diameter of 3 mm, outer diameter of 7 mm, and thickness of 2 mm). The microwave absorption measurements were carried out with an Anritsu MS4642A network analyzer over a frequency range of 2−18 GHz at room temperature. The RL values were calculated using the measured complex permittivity and permeability.
structure on the dielectric loss and magnetic loss of Fe3O4/ CNT nanocomposites. In this work, Fe3O4 compact-coated CNTs (FCCs) and Fe3O4 loose-coated CNTs (FLCs) with a fixed CNT/Fe3+ mass ratio of 1:2 are prepared, and the effects of the Fe3O4 coating structure on their microwave absorption behavior are investigated. The results show that the microwave absorption effectiveness, complex permittivity, and complex permeability of the Fe3O4/CNT composites are strongly correlated to the Fe3O4 coating structure. The minimum reflection loss (RLmin) value of the FCCs is much lower than that of the FLCs. Afterward, FCCs with CNT to Fe3+ mass ratios varying from 1:1 to 1:5 are prepared, and the microwave absorption properties are optimized. The lowest RL value of −43 dB and the widest absorption bandwidth range from 9.7 to 18 GHz for RL lower than −10 dB are achieved for the FCCs with a CNT to Fe3+ mass ratio of 1:4 (noted as FCC14) at thicknesses of 1.5 mm and 1.75 mm, respectively. Upon comparison with the literature related to Fe3O4-carbon-based composites, the FCC14 display the most effective specific RLmin and widest effective bandwidth (RL < −10 dB), indicating that the asprepared FCCs possess an excellent MAP.
3. RESULTS AND DISCUSSION 3.1. Microstructures and Properties of FCCs and FLCs. To obtain different coating structures, two representative FCC and FLC samples with the same CNT to Fe3+ mass ratio of 1:2 were prepared based on our previous works.33−35 The compact and loose coating structures were achieved by controlling the amount of alkali and solvent to tune the redox rate of Fe3+. Representative TEM images of FCCs and FLCs are shown in Figure 1a and Figure 1b. The Fe3O4 nanoparticle size in the FCCs is clearly smaller than in the FLCs. At the same time, the surface area covered by Fe3O4 and the number of Fe3O4 nanoparticles in the FCCs are obviously higher than in the FLCs. The schematic diagrams of the FCCs and FLCs are shown in the top left corners of Figure 1a and Figure 1b. For FCCs, the CNT surfaces are densely covered by small nanoparticles, while for FLCs, large and sparsely distributed nanoparticles are seen on the CNT surfaces. The EDS analysis (Figure S2) reveals that both FCCs and FLCs consist of the elements of carbon, oxygen, and iron. In addition, the weight ratio of C:O:Fe in the FCCs and FLCs is almost identical, indicating that the CNT to Fe3O4 mass ratio in the FCCs and FLCs is the same. Although the FCCs and FLCs are both uniformly coated with Fe3O4 nanoparticles, their structures are different: (1) the mean sizes of Fe3O4 nanoparticles in the FCCs and FLCs are 8.25 and 10.06 nm, respectively (Figures S3a and S3b), (2) the surface area covered by Fe3O4 in the FCCs is obviously higher than in the FLCs, and (3) the number of Fe3O4 nanoparticles in the FCCs is greater than in the FLCs. To reveal the relationship between the structure of the Fe3O4 coating and the properties of the CNT composites, the MAP of Fe3O4-coated CNTs was investigated. According to the transmit line theory,11 the RL values of the CNT composites are calculated using the following equations:
2. EXPERIMENTAL SECTION 2.1. Materials. Ferric acetylacetonate (Fe(acac)3), triethylene glycol (TEG), ethanol, and sodium hydroxide (NaOH) were purchased from Lan Yi Co. Ltd., Beijing, China. All chemicals are analytical grade and were directly used without further purification. Multiwalled CNTs were purchased from Chengdu Organic Chemicals Co. Ltd., Chinese Academy of Sciences. 2.2. Synthesis of Fe3O4-Coated CNTs. FCCs and FLCs were both synthesized through a solvothermal method. Briefly, in the preparation of FCC samples, 0.1 g of CNTs was suspended in 20 mL of TEG with stirring. Fe(acac)3 (0.2 g, CNT/Fe3+ mass ratio of 1:2) was introduced into the solution and sonicated for 15 min. Afterward, 2 mL of NaOH (6 M) was added into 10 mL of TEG with stirring. Then, the solution obtained was added to the CNT solution with another 15 min of sonication. The mixture was then transferred into a 50 mL Teflon-lined autoclave and heated at 200 °C for 2 h. The FLC samples were obtained with the same CNT/Fe3+ mass ratio. Asreceived CNTs (0.1 g) were suspended in 50 mL of TEG with stirring, followed by adding 0.2 g of Fe(acac)3. The mixture suspension was sonicated for 30 min and transferred into two 50 mL autoclaves and then heated at 200 °C for 4 h. The other five FCC samples with CNT to Fe3+ mass ratios of 1:1 (FCC11), 1:2 (FCC12), 1:3 (FCC13), 1:4 (FCC14), and 1:5 (FCC15) were prepared similarly. The products obtained were washed with ethanol several times to eliminate the surfactant and then washed to neutral with distilled water. Finally, they were separated from water by a magnet and subsequently dried in an oven (60−80 °C) for 12 h. The solvothermal method used for fabricating Fe3O4-coated CNTs is more stable and repeatable compared to the often used coprecipitation method because the synthesis parameters can be exactly controlled. To examine the stability of this method at large scale synthesis, the scale-up synthesized products were also obtained by five-time amplification. The SEM images (Figure S1 in Supporting Information) show the good stability of this coating method at large scale preparation of Fe3O4-coated CNTs since the coated CNTs are successfully obtained at the large scale. 2.3. Characterizations. The morphology of the Fe3O4-coated CNTs was observed by a transmission electron microscope (TEM, JEM2100, Japan) and a scanning electron microscope (SEM, Hitachi S-4800, Japan) in secondary electron scattering mode at 10 kV. High resolution transmission electron microscopy (HRTEM) images were taken using a JEOL JEM-2010 TEM at an accelerating voltage of 200 kV. The sizes of Fe3O4 nanoparticles were measured from the TEM images using the Nano Measure software. X-ray energy-dispersive
RL (dB) = 20 log|(Z in − 1)/(Z in + 1)|
(1)
Z in =
(2)
μr /εr tanh[j(2π /c) μr /εr fd]
where Zin is the normalized input impedance, f is the microwave frequency, c is the velocity of light, d is the thickness of the absorber, and εr and μr are the complex permittivity and permeability of the absorber medium, respectively. In our calculation, d values were variously 1, 1.5, 2, 2.5, and 3 mm. The RL values calculated for the FCCs and FLCs are shown in Figure 1c and Figure 1d, in which the RL values of the 2 mm thick sample were measured directly; the RL values corresponding to other thicknesses were calculated according to eq 2. It is obvious that the RL valley values shift to a lower 2975
DOI: 10.1021/acsami.6b13142 ACS Appl. Mater. Interfaces 2017, 9, 2973−2983
Research Article
ACS Applied Materials & Interfaces
within a certain frequency range, which is similar to the result for Fe3O4/SnO2 core/shell nanorods reported by Chen et al.37 To obtain excellent microwave absorption, high impedance matching is necessary, which should meet the needs of the complex permeability and permittivity equally.38 The difference between the complex permeability and permittivity of the FLCs is much larger than in the FCCs due to poor impedance matching. Consequently, the FLCs display weak microwave absorption behavior. This difference can be further demonstrated through the dielectric loss tangent (tan δe = ε″/ε′) and the magnetic loss tangent (tan δμ = μ″/μ′) as shown in Figure 1g and Figure 1h. Moreover, the FCCs and FLCs both have strong dielectric loss and weak magnetic loss over the whole frequency range, indicating that the RL is mainly attributable to dielectric loss rather than magnetic loss. Meanwhile, the magnetic loss tangents show severe fluctuations. 3.2. Difference in Absorption Mechanisms between the FCCs and FLCs. It is evident that the complex permittivity of Fe3O4-coated CNTs is dominated by the CNTs, while the magnetic response is mainly contributed by the Fe3 O4 nanoparticles. It is known that the excellent microwave absorption of composites is strongly dependent on the efficient complementarities between relative permittivity and permeability.37,39 The complex permittivity and permeability originate from various types of polarization and magnetic properties, which strongly depend on the size, crystallization, and structure of materials.11,27 As revealed in Figures S5 and S6, the size and crystallization of the Fe3O4 nanoparticles only slightly influence the microwave absorption behavior of the FCCs and FLCs. This implies the coating structure plays a crucial role in improving the microwave absorption effectiveness. The responsive behavior of Fe3O4-coated CNTs to incident electromagnetic waves is vital to their MAP. As depicted in Figure 2, when electromagnetic waves are incident into a material (red arrows), they can be reflected (blue arrows), absorbed (green arrows), or multireflected (continuous blue arrows). First, incident waves mainly transmit into composites
frequency with the increase of sample thickness, associated with quarter-wavelength attenuation.16 To be specific, the minimum RL values of FCCs with sample thicknesses of 1.5, 2, 2.5 and 3 mm are −22 dB (at 14.5 GHz), −16.87 dB (at 10.8 GHz), −18.62 dB (at 8.5 GHz), and −14.9 dB (at 6.7 GHz), respectively. Compared to the FCCs, as shown in Figure 1d, the RL of the FLCs is larger and the absorption bandwidths are broader. The RLmin value of the FLCs is −12 dB (at 8.7 GHz) for the sample with the thickness of 1.5 mm, which increases to −7.7 dB (at 3.7 GHz) at a thickness of 3 mm. Despite the CNT to Fe3+ mass ratio being the same for the FCCs and FLCs, their RL values are distinctly different. In addition, the RLmin value of the FLCs appears in a lower frequency range than that of the FCCs due to the Fe3O4 nanoparticle size difference between the FCCs and FLCs. According to the quarter-wavelength attenuation, the size of the Fe3O4 nanoparticles in the FLCs should be larger than in the FCCs, which is consistent with the Fe3O4 sizes measured (Figure S2a and Figure S2b). The bandwidth with effective absorption (BEA) is defined as the frequency range in which the RL < −10 dB, indicating that only 10% of the microwave energy is reflected while 90% is absorbed. As shown in Figure S4, the BEAs of FCCs with thicknesses of 1.0, 1.5, 2.0, 2.5, and 3.0 mm are 0, 3.4 ± 0.08, 2.2 ± 0.06, 1.75 ± 0.05, and 1.38 ± 0.08 GHz, respectively. However, the BEAs of FLCs are 0, 1.03 ± 0.07, 0.1 ± 0.02, 0, and 0 GHz, respectively, which is much lower than the BEAs of the FCCs. As mentioned above, the size and amount of Fe3O4 differ in the FCCs and FLCs; to investigate the size influence, absorption characteristics have been examined for the two kinds of pure Fe3O4 particles, and the results are given in Figure S5. The results show that only a tiny difference in the microwave absorption behavior can be detected between the two samples. Namely, compared to the mass ratio, the influence of the Fe3O4 particle size and quantity is negligible. Therefore, based on the same mass ratio, it can be concluded that the RL change largely arises from the difference of the Fe3O4 coating structure. To reveal the mechanism that gives rise to the RL difference between the FCCs and FLCs, the complex permittivity (εr = ε′ − jε″) and complex permeability (μr = μ′ − jμ″) of the FCCs and FLCs are compared. It is known that the real permittivity (ε′) and the real permeability (μ′) represent the ability of an absorber to store electric and magnetic energy, respectively, while the imaginary permittivity (ε″) and the imaginary permeability (μ″) represent the ability of an absorber to dissipate electric and magnetic energy, respectively.36 As shown in Figure 1e, the ε′ and ε″ values of the FCCs change slightly over the whole frequency range. For the FLCs, abrupt decreases from 2 to 6 GHz and from 2 to 8 GHz for ε′ and ε″ are seen in Figure 1f. In addition, both the ε′ and ε″ values of the FLCs show a small fluctuation within the frequency range of 6 to 18 GHz. Clearly, the ε′ and ε″ values of the FLCs are much higher than those of the FCCs. As revealed in Figure 1a and Figure 1b, the bare CNT surface area of the FLCs is larger than that of the FCCs, which results in both the electric polarization and electrical conductivity of the FLCs being higher than those of the FCCs. As shown in the orange curves of Figure 1e and Figure 1f, the μ′ values of the FCCs and FLCs are in the ranges of 0.99−1.37 and 0.98−1.37, respectively, whereas the μ″ values of the FCCs and FLCs are in the ranges of (−) 0.1−0.16 and (−) 0.07−0.17, respectively. The apparent fluctuations of the FLCs are attributed to the Fe3O4 particle size nonuniformity.36 Moreover, the μ″ values are negative
Figure 2. Microwave absorption mechanism of FCCs (a) and FLCs (b). 2976
DOI: 10.1021/acsami.6b13142 ACS Appl. Mater. Interfaces 2017, 9, 2973−2983
Research Article
ACS Applied Materials & Interfaces
Figure 3. SEM images of (a) pure CNTs and FCCs with various CNT to Fe3+ mass ratios: (b) 1:1 (FCC11), (c) 1:2 (FCC12), (d) 1:3 (FCC13), (e) 1:4 (FCC14), (f) 1:5 (FCC15). (g) RL curves of FCC11−FCC15 at 2−18 GHz (30 wt % in paraffin composite, 2 mm thickness). (h) Plots of attenuation constant vs frequency of FCC11−FCC15 samples.
microwave absorption. Consequently, composites with the magnetic particle compact-coated structure are ideal to achieve a high MAP attributed to the high incidence probability, multiple reflection and absorption of microwaves. 3.3. Optimization of the Coating Mass Ratio for Microwave Absorption Effectiveness Enhancement. The mass ratio of Fe3O4 to CNTs is another key factor determining the microwave absorption properties of Fe3O4-coated CNTs. To optimize their MAP, FCC samples with CNT to Fe3+ mass ratios changing from 1:1 to 1:5 (defined as FCC11−FCC15) were prepared. As shown in Figure 3a−f, the morphologies of pure CNTs and CNT nanocomposites were observed with SEM, which confirms the successful fabrication of Fe3O4 compact-coated CNTs. The surfaces of FCC11−FCC15 are all tightly and densely decorated with Fe3O4 nanoparticles, and no obvious aggregation could be detected. In addition, the amount of Fe3O4 nanoparticles decorated on the surfaces of CNTs increases as the CNT to Fe3+ mass ratio decreases. However, when the mass ratio of CNT to Fe3+ decreased to 1:5 (FCC15), redundant nanoparticles were seen. To investigate the microwave absorption properties of FCC11−FCC15, their RL values were calculated according to eq 1 and eq 2. As shown in Figure 3g, the RLmin values of FCC11−FCC15 in the frequency range of 2−18 GHz are −15, −17, −26, −37, and −27 dB, respectively, and the FCC14 sample shows the lowest RLmin values of the five samples. It is known that the performance of MAP reflects frequency, effective bandwidth, and RLmin values, so it is meaningful to investigate materials with good microwave absorption proper-
with less reflection occurring on the surface, which is attributed to the impedance matching between the wave absorbers and free space.40 As shown in Figure S7, most of the incident waves are reflected for the original CNTs. However, the reflection probability of the incident waves will decrease with the increase of the Fe3O4 nanoparticle coating area on the outer layer of the CNTs. In addition, small nanoparticles with large curvatures would induce multiple reflection, which is also beneficial for the incident waves transmitting into the composites.40 As indicated in the schematic, the incident waves transmitted into the absorbent in Figure 2a are more prevalent than that in Figure 2b due to its larger Fe3O4 nanoparticle coating area and smaller particle size. Second, space charge accumulation and interfacial polarization loss would arise at the interface of Fe3O4-coated CNTs, which is beneficial for improving microwave absorption. After multiple reflections, as indicated by the green arrows in Figure 2a and Figure 2b, the microwaves transmitted into the composites are dissipated into other kinds of energy, e.g., microwave absorption. Third, the continuous layer at the surface of CNTs, composed of closely spaced small nanoparticles, could amplify the response to incident microwaves through collective interfacial dipoles,40−42 thus improving the dielectric loss.43 In addition, the high surface area in Fe3O4coated CNTs provides more active sites than bare CNTs for dissipating and scattering microwaves.44 Finally, it is worth mentioning that a broad size distribution is unfavorable for achieving a high MAP.45 Importantly, the above factors should create more effective complementarities between the dielectric loss and the magnetic loss, which is vital for achieving excellent 2977
DOI: 10.1021/acsami.6b13142 ACS Appl. Mater. Interfaces 2017, 9, 2973−2983
Research Article
ACS Applied Materials & Interfaces
Figure 4. (a) Contour map of absolute values of the bandwidth (