Introduction

Electromagnetic interference (EMI) shielding of radio frequency radiation continues to be a serious concern in modern society due to increasing use of commercial, military and electronic devices. In order to protect the society from these harmful radiations, various shielding materials have been used in past. Conventional metals were the first choice for researchers and industrialists till the last few decades. Compared to conventional metal-based EMI shielding materials, electrically conducting polymer composites have gained popularity recently because of their light weight, resistance to corrosion, flexibility, and processing advantages (Liu et al. 2007). The EMI shielding effectiveness (SE) of a composite material depends mainly on the filler’s intrinsic conductivity, dielectric constant and aspect ratio (Li et al. 2006). The high conductivity, small diameter, high aspect ratio and superior mechanical properties of carbon nanotubes (CNTs) make them an excellent choice to be used as conductive composites for high performance EMI shielding materials. EMI shielding in the range of 8.2–12.4 GHz (X-band) is very important for military and commercial applications. Doppler, weather radar, TV picture transmission, and telephone microwave relay systems lie in this frequency range (Huang et al. 2007). Recently, several studies have been reported on CNTs reinforced thermoplastic polymer composites as effective and light weight EMI shielding materials in X-band. These include poly(methyl methacrylate) (PMMA) (Kim et al. 2004; Mathur et al. 2008b; Pande et al. 2009; Yuen et al. 2008), polystyrene(PS) (Mathur et al. 2008b; Yang & Gupta, 2005) polypropylene (PP) (Al-Saleh and Sundararaj 2009), polyurethane (PU) (Liu et al. 2007), polyvinylidene fluoride (PVDF) (Eswaraiah et al. 2011), poly(trimethyene terephthalate) (Gupta and Choudhary 2011), polyacrylate (Li et al. 2008), styrene acrylic emulsion (Li et al. 2010), cellulose triacetate (Basavaraja et al. 2011), ethylene vinyl acetate (EVA) (Das and Maiti 2008), reactive ethylene terepolymer (RET) (Park et al. 2010), and polycarbonate (PC) (Arjmand et al. 2011).

Epoxy resins are well established as thermosetting matrices for advanced structural composites, displaying a series of promising characteristics for a wide range of applications owing to their excellent mechanical properties, low cost, ease of processing, good adhesion to many substrates, and good chemical resistance (Garg et al. 2011). Several fibrous reinforcements e.g., carbon fibers, glass fibers, and aramid fiber (Kevlar fibers) etc. (Morais and Godfroid 2003) have been used with epoxy resin as matrix in producing structurally strong composite materials for commercial applications such as in aerospace industry. It suggests that CNT reinforced epoxy composites can also be a structurally strong EMI shielding material. Few studies have been reported on CNT–epoxy composites as EMI shielding material. Huang et al. (2007) have reported EMI-SE of 18 dB for composite with 15 wt% small single-walled carbon nanotubes (SWCNTs) and 23–28 dB for composite with 15 wt% long SWCNTs in the frequency band of 8–12.4 GHz. Li et al. (2006) observed the SE of 49.2 dB at 10 MHz for composite with 15 wt% long CNTs. It was reported that at higher frequencies (1 GHz), composites exhibited SE of 20 dB for both composites with 10 and 15 wt% SWCNT loadings. There are very few reports on CNT–epoxy composites for EMI shielding in X-band. Thus, there is a need to explore structurally strong MWCNT–epoxy composites for EMI shielding effectiveness in X-band. The reason behind fewer reports in CNT–epoxy composites is the need for high loading of CNTs required for high conductivity and EMI shielding. To improve the electrical conductivity and mechanical properties, higher loading of CNTs in EMI shielding composites is required (Park et al. 2009). But dispersion of high loading of CNTs in epoxy resin is difficult due to formation of agglomerates by the conventional techniques. However, epoxy composites synthesized using the conventional methods generally have low CNT contents. It has been reported that beyond 0.6 wt% of MWCNT, CNT tend to agglomerate (Yang et al. 2009) resulting in poor bending strength and modulus of the composites.

It is therefore important to develop a technique to incorporate higher CNT loading in epoxy resin without sacrificing their mechanical properties. Recently, several methods have been developed for fabricating CNT/polymer composites with high CNT loadings. One such technique is mechanical densification technique where vertically aligned CNTs were densified by the capillary-induced wetting with epoxy resin (Wardle et al. 2008).This technique is limited by the sample size. In another technique, a filtration system was used to impregnate the epoxy resin into CNT bucky paper (Gou 2006; Wang et al. 2004). However, it was very difficult to completely impregnate the bucky paper with epoxy resin. Recently, Feng et al. (2010) has reported a mixed curing-assisted layer-by-layer method to synthesize MWCNT/epoxy composite film with high CNT loading from ~15 to ~36 wt%. The electrical conductivity of the composites showed a value of 0.12 S/cm. However, the mechanical properties of the composites were not reported. In another study by Feng et al (2011), up to ~39.1 wt% SWCNT–epoxy composites were fabricated having enhanced mechanical properties. However, in all the above studies, EMI shielding effectiveness of these composites has not been reported. In one of our previous study, we observe that it was not possible to disperse more than 0.5 % by weight of MWCNT in the epoxy resin due to agglomeration of the tubes at higher loadings (Garg et al. 2011). These small loadings did not contribute significantly toward EMI shielding despite having high electrical conductivity of these tubes. Therefore, a new technique was used to increase the loading of MWCNT by coating the individual nanotube with epoxy resin in the form of prepreg. We report here the electrical, mechanical, thermal, and EMI shielding of such composites. Scanning electron microscope (SEM) and high resolution transmission electron microscope (HRTEM) are used to investigate morphological and microstructural properties to establish a correlation between microstructure and mechanical/electrical properties of the composites.

Experimental

Materials

Di-glycidyl ether of bisphenol A (DGEBA) type epoxy resin (LY-556, Huntsman Co. Inc.) was used as the matrix material. Aradur (HY5200, Huntsman Co. Inc.) was used as a curing agent. MWCNTs were synthesized using toluene as a carbon source and ferrocene as catalyst precursor in a CVD set-up established in the laboratory (Mathur et al. 2008a). The MWCNTs produced were 10–70 nm in diameter and ~200-μm in length (Singh et al. 2011). These were 90 % pure with 10 wt% of Fe catalyst.

Fabrication of MWCNT–epoxy composites

Fabrication process of MWCNTs–epoxy composites is depicted schematically in Fig. 1. The different amounts of as-grown MWCNTs (a) was dispersed in acetone for 2 h in ultrasonic bath for opening the bundles (b).The epoxy resin (c) was diluted with acetone to reduce the viscosity of the resin (d).The dispersed MWCNTs were added in diluted epoxy resin and magnetically stirred for 24 h to obtain a uniform dispersion (e).The curing agent was added in the ratio of 100 (epoxy) : 23 (hardener) by weight in the dispersed material and magnetically stirred for 30 min (f). The dispersed MWCNTs in epoxy resin were filtered in a specially designed filtration unit (g) to form a film of epoxy impregnated MWCNT and dried at 80 °C for 2 h to form MWCNT–epoxy prepreg (h). The prepreg was compression molded in hydraulic press between two plates at 80 °C followed by curing under press at 150 °C for 4 h. The resultant composite paper containing different percentage of CNTs were obtained in the form of a uniform circular disk of 10 cm dia. This compression molded cured film was cut into the desired shape (i) for further testing. For the preparation of composite block (k) from these prepreg, the prepreg papers were cut into the shape of 60 × 20 mm2 size (j) and placed in a three piece mold for compression molding in hydraulic press.

Fig. 1
figure 1

Schematic diagram for preparation of high loading MWCNT–epoxy composites by filtration followed by compression molding

Full size image

Characterization

The fractured surface of MWCNT–epoxy composite samples was analyzed by SEM (Leo model: S-440). HRTEM studies of MWCNTs were carried out using Technai G20-stwin, 300 kV instrument. The MWCNT content in epoxy matrix was determined using a thermogravimetric analyzer (TGA) (Mettler Toledo TGA/SDTA 851 e). The test was performed between 30 and 600 °C at a heating rate of 10 °C/min under nitrogen with a flow rate of 50 cc/min. Thermal diffusivity of the composite samples was measured using a Nflash Line 3000 (Anter make) system using a standard test method for thermal diffusivity by flash technique. The thermal conductivity of the samples was determined by measuring the specific heat of the sample on the same instrument. The measurement is based on ASTM-E-1461. The tensile strength and Young’s modulus of pure epoxy and MWCNT–epoxy composite film was measured using an Instron machine model 4411. The composite films were cut into standard dog bone shape (Mathur et al. 2008b; Allaoui et al. 2002) using ASTM D638. A special die-punch was used for the purpose. The gauge length and width of the test sample were 30 and 6 mm, respectively. The cross–head speed was maintained at 0.5 mm/min. The electrical conductivity of composite films (60 mm × 20 mm × 0.35 mm) was measured by d.c four probe contact method (Singh et al. 2011; Singh et al. 2008) using a Keithley 224 programmable current source for providing current using ASTM C611-98. The voltage drop was measured by Keithley 197 autoranging digital microvoltmeter. The values reported in the text were averaged over five readings of voltage drops at different positions of the samples. EMI-SE was measured by placing the composite film (0.35 mm ± 0.05 thick) inside X-band waveguides using a vector network analyzer (VNA) (E8263B Agilent Technologies). Magnetic measurements were performed using the vibrating sample magnetometer (VSM) model 7304 Lakeshore Cryotronics Inc., USA.

Results and discussion

Determination of weight percentage of MWCNT in composite

The weight fraction of CNT in each composite was evaluated using TGA. The TGA curves of epoxy, CNT, and composites are presented in Fig. 2. The weight loss was measured at 600 °C because thermal decomposition of MWCNT starts slightly above 600 °C. The percentages of CNTs in the epoxy obtained were 4.2, 15.1, and 20.4 wt%. The calculations were made using the method developed by Ogasawara et al. (Ogasawara et al. 2011). These were designated as CMP1, CMP2, and CMP3, respectively. Compression molded prepreg papers of CMP3 were designated as CMP4.

Fig. 2
figure 2

Thermogravimetric analysis results of CNT, epoxy, and CNT/epoxy composites

Full size image

Mechanical properties of the composite film

Figure 3a, b show the variation in Young’s modulus and tensile strength with MWCNT loading in epoxy composites. Young’s modulus increased with MWCNT loading and reached up to a value of 3203 MPa (CMP3) from 1990 MPa (neat epoxy). This is due to the addition of stiffer material (CNT) into epoxy resin. The tensile strength also increased with increase in MWCNT from 42 MPa (neat epoxy) to 59 MPa (CMP3), an improvement of ~40 % over the neat polymer. The SEM micrograph of fractured surfaces of the MWCNT–epoxy composite films after tensile testing (Fig. 4b–d) reveal that CNTs are uniformly dispersed without any visible aggregation. Aggregation can reduce the reinforcing effect of CNTs leading to a reduction in the strength of composites. A uniform dispersion of CNTs results in good load transfer from matrix to the CNTs resulting in improved mechanical properties at high loading of CNTs. This technique of formation of CNT-dispersed epoxy composites is more effective than conventional techniques of dispersion. This technique is also different than other epoxy impregnated bucky paper based technique where epoxy resin is impregnated into bucky paper. In those techniques, proper impregnation of epoxy is very difficult due to availability of very small pores between the CNTs. In the present technique, epoxy impregnated prepreg paper is prepared by dispersing long MWCNT into epoxy resin where proper adhesion takes place between the matrix and CNTs.

Fig. 3
figure 3

Variation in a tensile strength and b Young’s modulus with MWCNT loading

Full size image
Fig. 4
figure 4

Fracture surface of a pure epoxy, b 4.2 wt% MWCNT–epoxy, c 15.13 wt% MWCNT–epoxy and d 20.4 wt% MWCNT–epoxy composite film

Full size image

Electrical conductivity of the composites

Figure 5a shows the variation of room temperature electrical conductivity with the change in MWCNT loadings. Pure epoxy is electrical insulator with d.c. electrical conductivity value of ~10−15 S/cm (Barrau et al. 2003). However, conductivity of MWCNT–epoxy composites displays a systematic increase with the increase in MWCNT content from 4.2 to 20.4 wt%. The above enhancement can be attributed to the formation of extensive 3D networks of MWCNTs within epoxy matrix. In particular, the electrical conductivity is dramatically enhanced by 15 orders of magnitude for CMP1 compared to that of pristine epoxy. This can be attributed to combined effect of high aspect ratio of MWCNTs, prefect graphitic structure (interlayer spacing 0.342 nm, see HRTEM image, Fig. 5b) and its effective dispersion (Fig. 4b–d) within epoxy matrix. When the CNT loading is further increased to 20.4 wt% (CMP3), the electrical conductivity increased to 9 S/cm. The achieved conductivity value was near the prescribed range for microwave shielding. Therefore, good EMI shielding response is expected (Saini et al. 2012; Saini and Arora 2012).

Fig. 5
figure 5

a Variation of electrical conductivity of MWCNT–epoxy nanocomposites as a function of MWCNT content, b HRTEM image of individual MWCNT showing multiwall structure and well-defined lattice fringes with interplanar spacing of 0.342 nm

Full size image

Magnetization studies

Figure 6a shows the vibrating sample magnetometer (VSM) plot of CMP2, CMP3 and pure polymer whereas inset of Fig. 6a shows that VSM of MWCNTs and the magnetization data are presented in Table 1. It can be seen that MWCNT displays moderate ferromagnetic properties with saturation magnetization (M s), retentivity (M r) and coercivity (H c) values of ~7 emu/g, ~2.2 emu/g, and 577 G, respectively. The observed hysteresis loop can be attributed to the presence of iron particles inside the cavity of MWCNTs as shown in the HRTEM image (Fig. 6b). It is also seen that incorporation of above MWCNTs inside non-magnetic epoxy matrix results in introduction of magnetic properties. Therefore, the M s, M r, and H c values were found to be 0.79 emu/g, 0.21 emu/g, and 480 G, respectively for CMP2 and 1.33 emu/g, 0.46 emu/g, and 570 G, respectively, for CMP3. These results are of specific interest as good magnetic properties are important in modulating the shielding response (Saini et al. 2012).

Fig. 6
figure 6

a Magnetization studies of pure epoxy, CMP2 and CMP3 and inset shows for pure MWCNTs showing ferromagnetic character and hysteresis loop. b HRTEM image of individual MWCNT showing entrapped iron-phase within internal cavity of tubes, inset shows zoom-in image of entrapped iron

Full size image
Table 1 Magnetization data for various samples
Full size table

EMI shielding performance

The EMI shielding is a direct consequence of reflection, absorption, and multiple internal reflection losses at the existing interfaces, suffered by incident electromagnetic (EM) waves. The total shielding effectiveness (SET) can be expressed as (Saini et al. 2009; Singh et al. 2012):

$$ {\text{SE}}_{\text{T}} \left( {\text{dB}} \right) = 10\,{ \log }_{ 10} ({\text{P}}_{\text{T}} /{\text{P}}_{\text{I}} ) = 20\,{ \log }_{ 10} \left( {{\text{E}}_{\text{T}} /{\text{E}}_{\text{I}} } \right) = { \log }_{ 10} \left( {{\text{H}}_{\text{T}} / 20{\text{H}}_{\text{I}} } \right) $$
(1)

where P I (E I or H I) and P T (E T or H T) are the power (electric or magnetic field intensity) of incident and transmitted EM waves, respectively. The scattering parameters S 11 (S 22) and S 12 (S 21) of VNA are related to reflectance (R) and transmittance (T), respectively, i.e., T = |E T/E I|2 = |S 12|2 (=|S 12|2), R = |E R/E I|2 = |S 11|2 (=|S 22|2). Therefore, attenuations due to reflection (SER) and absorption (SEA) can be conveniently expressed as:

$$ {\text{SE}}_{\text{R}} { = 1}0\,{ \log }_{ 10}\,\left( { 1- {\text{R}}} \right) $$
(2)
$$ {\text{SE}}_{\text{A}} = 10\,{ \log }\,\left( { 1- A_{\text{eff}} } \right) = 10\,{ \log }_{ 10} \left[ {T/\left( {1 - R} \right)} \right] $$
(3)

Further, the angular frequency (\( \omega \)) dependence of reflection and absorption losses can be expressed in the terms of total conductivity (\( \sigma_{T} \)), real permeability (\( \mu{'} \)), skin depth (\( \delta \) and thickness (t) of the shield material as (Singh et al. 2011):

$$ {\text{SE}}_{\text{R}} \left( {\text{dB}} \right) = - 10\;\log_{10} \left( {\frac{\sigma_{T}}{{16\omega \varepsilon o\mu{'} }}} \right) $$
(4)
$$ {\text{SE}}_{\text{A}} \left( {\text{dB}} \right) = - 20\frac{t}{\delta }\log_{10} \;e = - 8.68\left( {\frac{t}{\delta }} \right) = - 8.68t\left( {\frac{{\sigma_{T}\omega \mu{'} }}{2}} \right)^{\frac{1}{2}} $$
(5)

The above equations reveals that both conductivity as well as magnetic properties are useful for enhancing the absorption and hence total shielding. Therefore, use of MWCNT as conducting filler with additional magnetic properties (due to entrapped iron particles) is expected to enhance shielding effectiveness.

The shielding effectiveness value as well as electromagnetic attributes (permittivity and loss tangent) of samples is shown in Fig. 7. The results reveal that pure epoxy sample gives negligible attenuation (not shown) with (SET ≈ −0.3 dB). However, with the addition of MWCNT, both reflection (SER) as well as absorption (SEA) loss component increases resulting in the enhancement of total shielding (SET) i.e., from −10 (for CMP1) to −19 dB (CMP3) as shown in Fig. 7a. Further, the careful analysis of underlying reflection and absorption components show that SEA/SER ratio (Fig. 7b) increases with incorporation of higher amount of CNTs. These results demonstrate that SET is dominated by absorption component, contribution of which increases with increasing MWCNT content.

Fig. 7
figure 7

Frequency dependence of a total shielding effectiveness (SET), b SEA/SER, c real permittivity (ε′) and d loss tangent (tan δe), for MWCNT–epoxy nanocomposites

Full size image

The permittivity spectra of these samples show that real permittivity (dielectric constant) value decreases with increase of frequency. This can be ascribed to the inability of polarization vector to maintain in-phase movement with incident high frequency electromagnetic radiation. Further, the dielectric constant increases with loading of higher amount of CNTs which may be ascribed to the Maxwell–Wegner interfacial polarization. The large difference in the electrical conductivity of MWCNT (~104 S/cm) filler and epoxy matrix (10−15 S/cm) resulted in charge localization at the interfaces leading to polarization and related losses. In order to explore the reason behind the enhanced absorption loss (SEA), the dielectric loss tangent values (i.e., tan δe = \( {{\varepsilon{''} } \mathord{\left/ {\vphantom {{\varepsilon{''} } {\varepsilon{'} }}} \right. \kern-0pt} {\varepsilon{'} }} \)) have also been calculated which reflect the ability of a material to convert applied energy into heat. Therefore, materials with high tan δe value are useful for making microwave absorbing materials in stealth technology. The results have shown that tan δe increases from CMP1 to for CMP3 which is responsible for enhanced shielding response.

Figure 8a shows the magnetic permeability of the composites which shows weak frequency dependence. Furthermore, the permeability was found to increase with CNT loading which is due to the presence of iron particles within the cavity of CNT. The magnetic properties help in better matching of input impedance, reduction of skin depth, and additional magnetic losses. Therefore, they contribute toward improvement of absorption loss.

Fig. 8
figure 8

a Frequency dependence of real magnetic permeability (μ′) and b temperature dependence of thermal conductivity (k) of the MWCNT–epoxy shield

Full size image

The above dielectric and magnetic losses are responsible for generation of heat due to molecular friction. The safe and fast dissipation of this heat is desirable for maintaining long term and stable performance of the shield. Like other polymers, pure epoxy matrix also has very low thermal conductivity ~0.14 W/mK (Cui et al. 2011) and displays poor heat dissipation response. The addition of MWCNT with high inherent thermal conductivity (k ~ 3000 W/mK) leads to thermal conductivity value (Fig. 8b) of 2.175 W/mK for CMP4. Such an improvement (~15 times higher than pure epoxy) in k gives direct evidence of improvement of microwave heat dissipation characteristics of the shield. The variation in the thermal conductivity with temperature is very small (from 2.0 W/mK at 50 °C to 2.175 W/mK at 200 °C) and increases slightly with temperature. Thermal conductivity of any composite material depends mainly on the filler conductivity, dispersion, orientation, and the interfacial thermal resistance between filler and polymer. The change in thermal conductivity with temperature depends mainly on the interfacial thermal resistance which decreases with increase in temperature (Jakubinek et al. 2010). This results in slightly higher thermal conductivity at higher temperature.

As we increase the shield thickness by stacking of layers, geometrical effects tend to come into picture and a clear loss peak is observed in the shielding curve. Moreover, as the shield thickness increases, the loss peak intensity increases and shifts toward lower frequency. This may be due to the combined effect of dielectric and magnetic properties along with geometrical (thickness) effects resulting in better matching of input impedance. Therefore, a five-layered shield with total thickness of 1.75 mm resulted in SET value of −60 dB at 8.3 GHz (Fig. 9a) which corresponds to blocking of more than 99.999 % of incident EM radiation. To prove further, SET was divided into reflection and absorption components as shown in Fig. 9b which clearly demonstrates the dominance of absorption. In order to explore the superiority of layer stacking method, a block of above composite (CMP4) with thickness of ~1.75 mm was also prepared by compression molding. The shielding response of the block (Fig. 9c) revealed that maximum attenuation was −50 dB which was less than that observed for layered sample. Further, unlike layered sample no band was observed for bulk sample. To investigate the probable reasons, the SEA/SER ratio was calculated (Fig. 9d) for a bulk and layered samples of same thickness i.e., 1.05 mm (equivalent to three layers). The results revealed that in the entire frequency range, the ratio was higher for layered sample than for its bulk counterpart. This reflected that for layered sample absorption was playing more significant role than the bulk sample. The same was also complemented by the fact that ratio of absorbance value of the film and bulk sample (inset Fig. 9d) was greater throughout the entire frequency range. Furthermore, reflection loss of layered sample (~−7.0 dB, Fig. 9b) was much less than block sample (~−10 dB, Fig. 9c) which signifies better impedance matching for the former. The better impedance matching along with higher absorption capability of the layered film sample resulted in attainment of band and higher shielding effectiveness value. The above results implied that layering is a better method of obtaining higher shielding effectiveness compared to compression molded thick block of same thickness. The observed attenuation herein crossed the limit (−30 to −40 dB) of commercial applications, which suggests that these nanocomposites are promising candidates for making futuristic radar absorbers. These results were also compared with recently conducted research on CNT–polymer composites in X-band and tabulated in Table 2. This table shows that the present result is superior to reported results.

Fig. 9
figure 9

Frequency dependence of a SET of CMP3 with no. of layers. b SET, SEA, and SER of five layers of CMP3. c SET, SEA, and SER of CMP4 of 1.75 mm thickness. d SEA/SER for layered sample and CMP4 at 1.05 mm thickness, inset shows the absorption ratio of layered film vs compression molded block

Full size image
Table 2 Results of shielding effectiveness of CNT based various polymer composites reported by previous authors in X-band
Full size table

Conclusions

MWCNT–epoxy composites were fabricated using a novel dispersion and compression molding technique which enabled dispersion of high loadings of CNTs (up to 20.4 wt%) uniformly. A value of 9 S/cm of electrical conductivity was achieved which helped in attaining high EMI shielding properties of such composites i.e., up to 60 dB in X-band which is the most desired range for commercial applications such as radar, TV picture transmission, and telephone microwave relay systems, etc. An addition of MWCNT in the epoxy also provided structural integrity to the composites with tensile strength of the order of 60 MPa along with improved thermal conductivity which is a prerequisite for efficient heat dissipation in microelectronics devices.