Critical insights into understanding the effects of synthesis temperature and nitrogen doping towards charge storage capability and microwave shielding in nitrogen-doped carbon nanotube/polymer nanocomposites

Abstract

In this study, various nitrogen-doped (N-doped) multiwall carbon nanotubes (MWNTs) were synthesized by varying the synthesis temperature (650 °C, 750 °C and 850 °C), and their charge storage capability and electromagnetic (EM) shielding effectiveness (SE) were assessed by incorporating them into a PVDF (polyvinylidene fluoride) matrix. Nitrogen doping was adopted to generate numerous polarizable centers in MWNTs. The concentration of nitrogen and polarizing centers was optimized by varying the synthesis temperature. The nitrogen doping had a significant impact on the structural, thermal, and electrical properties of MWNTs. Dielectric spectroscopy of the nanocomposites containing self-polarizable MWNTs showed significantly low loss tangent, exhibiting good charge storage ability at a given concentration of MWNTs. The electrical conductivity of N-doped nanocomposites decreased as the synthesis temperature increased from 650 °C to 850 °C. This phenomenon was observed to be significantly different to the bulk powder. The electrical conductivity of the nanocomposites was also reflected in the EM shielding results where the nanocomposites containing N-doped MWNTs showed lower shielding effectiveness than the un-doped MWNTs. Moreover, the SE decreased with increasing synthesis temperature for N-doped MWNTs. Taken together, this study demonstrates critical insights about the impact of nitrogen doping and synthesis temperature on electrical conductivity, charge storage ability, and EM shielding of MWNT polymer-based nanocomposites.

---

1. Introduction

The recent development in multifunctional nanocomposites is paving the way for numerous applications in the modern electronics and communication sector. In the recent past, polymer-based nanocomposites have gained critical attention as high-performance multifunctional materials because they not only manifest the properties of their constituents but also depict synergetic performance through interactions of individual blend constituents. In this context, various lightweight polymer-based nanocomposites were developed containing different inclusions, such as carbonaceous, metallic, and ceramic nanoparticles, for specific applications.^1–3

More recently, electromagnetic interference (EMI) shielding materials have gained enormous attention not only due to their vital role in shielding microwave radiation in military sector but also to control the electromagnetic pollution. Metals are classical microwave shielding materials but reflection is their primary mechanism of shielding, thereby failing to control the EM pollution. Moreover, high density and susceptibility towards corrosion restrain their applications further. Since absorption is the desirable way to avoid EM interference and to control the EM pollution, the research to develop EMI shielding materials is mainly concentrated towards polymer-based nanocomposites containing various microwave-active nanofillers due to their efficient microwave-absorption ability.^4–6

It has been reported that materials with high microwave-shielding ability should possess desirable electrical conductivity and electric and magnetic dipoles.⁷ Since most polymers are insulating, the dispersion of electrically conducting nanofillers becomes an effective practice to tailor the electrical properties of the nanocomposites. In this context, owing to its high aspect ratio, carbon nanotube (CNT) has become prime constituent, as a conducting inclusion, to achieve high electrical conductivities at low concentrations. In case of polymeric nanocomposites containing CNTs, the microwave shielding arises due to ohmic and polarization losses. The percolated network of CNTs in the host matrix leads to conducting paths for nomadic charges, contributing to large ohmic losses.^8,9 Over the years, it was realized that higher concentration of CNTs leads to higher ohmic losses and subsequently higher microwave attenuation.^6,10–12 It has been well established that the 1D nanofillers has exceptional mechanical properties.^13–16 However, high nanofiller concentration often leads to poor mechanical properties in the nanocomposites and hence are economically non-viable.^17,18

Herein, the rational was to achieve appreciable microwave attenuation at low concentration of MWNTs. A unique approach was adopted to achieve self-polarizable MWNTs through nitrogen-doping (N-doping). The incorporation of pyridinic, pyrrolic, and quaternary nitrogen leads to multiple polarizing centers, which could be utilized for microwave attenuation through dielectric polarization. However, nitrogen doping leads to structural defects on MWNTs, which can disturb nomadic charges transport, and therefore, severely reduce the ohmic loss.¹⁹ Hence, the balance between ohmic loss and polarization loss is the key point to achieve synergistic microwave attenuation. Based on the literature, though various MWNT modifications were studied for EMI shielding;^20–24 nonetheless, detailed microwave-attenuation properties of self-polarizable MWNTs synthesized by nitrogen doping are not well understood.

In this study, self-polarizable MWNTs were synthesized using nitrogen doping. The polarizing centers were systematically varied by synthesizing nitrogen-doped (N-doped) MWNTs at various temperatures, i.e. 650 °C, 750 °C and 850 °C. Detailed characterizations like aspect ratio, structural defects, presence of nitrogen, and electrical conductivity of synthesized MWNTs were carried out. In order to realize the self-polarization effect of MWNTs, un-doped MWNTs were synthesized at 650 °C and their properties were compared with N-doped MWNTs. The dielectric and EMI shielding properties of MWNTs were analyzed by dispersing them in a PVDF matrix by melt extrusion process. In brief, this study aims at bringing out critical insights in understanding the effect of N-doping and MWNTs synthesis temperature towards dielectric properties and EM attenuation in PVDF-based nanocomposites.

2. Experimental

The materials used in this study, synthesis of un-doped and N-doped MWNTs, nanocomposites preparation, MWNTs and nanocomposites characterization can be found in the ESI.†

3. Results and discussion

3.1. MWNTs characterizations

The structure and aspect ratio of MWNTs play an important role in charge transport properties and attenuation of EM radiation.²⁵ Moreover, the percolation of MWNTs in host matrix is governed by aspect ratio and it has been well reported that the percolation threshold decreases with increasing aspect ratio.^26–29 Hence, overall morphology and graphitic structure of MWNTs were assessed at low and high magnifications using TEM micrographs, respectively. Fig. 1 shows TEM micrographs of various types of MWNTs and the inset depicts high-magnification micrographs for the respective MWNTs. One can easily realize the striking effect of nitrogen doping on the structure of MWNTs. The un-doped MWNTs manifested an open-channel structure, whereas the N-doped MWNTs revealed bamboo-like structure.^30–36

Fig. 1 TEM micrograph of (a) un-doped MWNTs; TEM micrographs of N-doped MWNTs synthesized at (b) 650 °C, (c) 750 °C, and (d) 850 °C. Insets show high-magnification TEM micrographs of respective MWNTs.

In this study, since the catalyst used for the synthesis of both un-doped and N-doped MWNTs was similar, the significant difference observed in the structure was exclusively due to nitrogen doping. Multiple walls of MWNTs are also well evident from high-magnification TEM micrographs. In case of N-doped MWNTs, the growth of outer walls of MWNTs is straight and continuous, whereas inner walls were severely distorted and closed inside. Similar structure was evident for all N-doped MWNTs synthesized at different temperatures. This is well correlated to local distortion of graphitic structure of N-doped MWNTs by substitution of nitrogen, resulting in less ordered structure.

Interestingly, the un-doped MWNTs had straight, open channels with highly ordered and less defective structure. In recent past, it was observed that the N-doped MWNTs synthesized using Co catalyst manifested an open-channel structure, whereas MWNTs grown on Fe catalyst showed a bamboo-like structure.³⁷ The graphitic envelope can be easily formed around Fe compared to Co; this facilitates the bamboo-like growth of MWNTs over the Fe catalyst. Therefore, we can conclude that the bamboo-like structure is not only due to nitrogen doping effect but also because of Fe catalyst. Hence, the synergistic effect of Fe catalyst and nitrogen doping leads to the bamboo-like structure of MWNTs.^38–41

Table 1 shows average diameter, length, aspect ratio, and catalyst size of various types of synthesized MWNTs. The un-doped MWNTs manifested highest aspect ratio of 113, whereas nitrogen doping resulted in low-aspect-ratio MWNTs. As reported earlier^42–44 and verified in our case too, the diameter of MWNTs was observed to be in well correlation with catalyst size. Interestingly, it was observed that the diameter of MWNTs significantly increased with the synthesis temperature. This corresponds to the temperature-sensitive events like sintering, liquefaction, and coalescence of Fe nanoparticles at higher temperature, resulting in larger size of catalyst nanoparticles.

Table 1 Average length, diameter, aspect ratio, and catalyst size data for un-doped MWNTs and N-doped MWNTs synthesized at various temperatures

	Length (nm)	Diameter (nm)	Aspect ratio	Cat. size (nm)
Un-doped (at 650 °C)	1700 ± 17	15 ± 0.7	113	17
N-Doped (at 650 °C)	1266 ± 12	50 ± 2.5	26	40
N-Doped (at 750 °C)	2564 ± 25	46 ± 2.3	56	50
N-Doped (at 850 °C)	2787 ± 27	64 ± 3.2	44	55

---

At 650 °C, the un-doped MWNTs were significantly longer than the N-doped MWNTs. This can be ascribed to the open-channel structure of un-doped MWNTs where all carbon atoms contributed to the axial growth of MWNTs, whereas in case of N-doped MWNTs the axial growth was minimum due to bamboo-like structure. However, it is interesting to note that the N-doped MWNTs synthesized at higher temperatures, i.e. 750 °C and 850 °C, were relatively long. At higher temperatures, the decomposition of gases is significantly higher, causing faster axial growth of MWNTs. According to these observations, it can be concluded that at a particular temperature the axial growth of MWNTs will be dissimilar for various precursor gases. N-Doped MWNTs synthesized at 850 °C showed larger diameter than 750 °C; however, their lengths were comparable. This resulted in smaller aspect ratio at 850 °C.

The presence of different types of nitrogen bonding has a great effect on structure and consequently electronic properties of MWNTs. Hence, in order to analyze the presence and concentration of nitrogen, elemental mapping was carried out using X-ray photoelectron spectroscopy (XPS). The total concentration of nitrogen was estimated by comparing the area ratios of N1s and C1s spectra. Basically, quaternary, pyrrolic, and pyridinic types of nitrogen were identified in N-doped MWNTs.^45–47 Furthermore, fourth type of nitrogen was recorded namely NO_x, corresponding to the oxidized state of pyridinic nitrogen or presence of intercalated N₂ in MWNTs.^38,39 Fig. 2a depicts contribution of various nitrogen substitutions to total nitrogen present in N-doped MWNTs. The N-doped MWNTs synthesized at low temperature, i.e. 650 °C, manifested highest concentration of nitrogen (3.85 at%).

Fig. 2 (a) % nitrogen content, (b) I_D/I_G ratio, (c) powder electrical conductivity, and (d) TGA plot for various synthesized MWNTs.

The effect of synthesis temperature on the nitrogen content is well evident. The nitrogen content decreased with increasing the synthesis temperature. The N-doped MWNTs synthesized at higher temperatures, owing to their highly ordered structure, as will be confirmed by Raman spectroscopy and TGA, lacked sites for hetero-atom substitution. Along with total concentration of nitrogen, synthesis temperature depicted striking impact on type of nitrogen bonding. In this context, fraction of quaternary and pyrrolic nitrogen was realized to increase with the synthesis temperature. In recent past, various studies asserted that the presence of nitrogen leads to highly defective structure, whereas the type of nitrogen bonding greatly affects the electronic properties of N-doped MWNTs. In this frame of reference, since quaternary and pyridinic types of nitrogen bonding are present in the MWNTs walls, they play a vital role in controlling the electronic properties of MWNTs compared to pyrrolic type of nitrogen bonding, which is only present at the edge of MWNTs.^37,48,49 Among all the nitrogen substitutions, pyridinic nitrogen is expected to result in higher amount of structural defects in MWNTs.^37,50 Pyridinic nitrogen as scattering center and its sequent defects can act as polarizing centers, influencing dielectric and subsequently microwave attenuation properties of nanocomposites.

The structural defects within synthesized MWNTs were assessed using Raman spectroscopy. In case of N-doped MWNTs, the defects generated due to nitrogen doping provide direct evidence of self-polarizability of MWNTs. The presence of nitrogen distorts hexagonal carbon structure, thereby creating a defective structure. Raman spectroscopy is a sensitive tool to understand the graphitic and defective structure of MWNTs. The G-band, which arises at 1600 cm⁻¹, corresponds to the characteristic C–C stretching of graphitic structures,^51,52 whereas the D-band (1400 cm⁻¹) emerges due to elastic and inelastic double-scattering event, denoting defective sites.⁵³

At large, pentagon–heptagon pairs, vacancies, heteroatoms, and impurities give rise to the D-band. For relative comparison, the intensity of the G-band is often used to normalize the intensity of the other bands. Fig. 2b depicts the intensity ratio of the D- and G-bands for various types of MWNTs. The un-doped MWNTs showed I_D/I_G ratio of 0.65, whereas N-doped MWNTs synthesized at similar temperature, i.e. 650 °C, depicted strikingly higher I_D/I_G ratio (0.79), manifesting highly defective structure. This is in line with the TEM micrographs where N-doped MWNTs were observed to be structurally inferior to un-doped MWNTs. This is mainly due to the incorporation of nitrogen in MWNTs structure, manifesting more defective sites.

Interestingly, I_D/I_G ratio for N-doped MWNTs decreased with the synthesis temperature, implying superior crystalline structure. In case of N-doped MWNTs, the amount of defects (I_D/I_G ratio) was in good agreement with the concentration of nitrogen. In fact, the presence of nitrogen generates defect in graphitic carbonaceous structures; therefore, increasing the concentration of nitrogen manifested higher I_D/I_G ratio. Similar observations indicating direct relationship between concentration of nitrogen and amount of defects have been well reported.^54–56 It was also captivatingly observed that the N-doped MWNTs synthesized at 850 °C presented less number of defects than un-doped MWNTs.

It is well accepted that the intrinsic conductivity of nanoparticles plays a vital role in deciding the final electrical properties of nanocomposites. Moreover, microwave attenuation through both reflection and absorption is governed by the electrical conductivity of dispersed nanoparticles.⁷ Hence, intrinsic electrical conductivity was evaluated by measuring the powder conductivity of un-doped and various N-doped MWNTs. Fig. 2c depicts the electrical conductivity of the compressed MWNTs powders. For instance, un-doped MWNTs manifested electrical conductivity of 17.5 S cm⁻¹. The N-doped MWNTs synthesized at the identical temperature showed slightly decreased electrical conductivity. This corresponds to the highly distorted structure of N-doped MWNTs. However, further increase in temperature resulted in increased conductivity where N-doped MWNTs synthesized at 850 °C manifested highest electrical conductivity of 35.7 S cm⁻¹. The conductivity of MWNTs synthesized at 850 °C was two folds higher than that of un-doped MWNTs. This is in well agreement with the structural properties of MWNTs where lower amount of defects and higher length of MWNTs paved way for effective nomadic charge transfer. The higher intrinsic conductivity can facilitate the attenuation of EM radiation by increasing both reflection and absorption. The effect of intrinsic conductivity on microwave absorption will be discussed in detail later on.

Thermal stability of the as-synthesized MWNTs was inspected using thermogravimetric analysis. All the experiments were carried out in air atmosphere. Herein, yield and thermal stability was assessed for un-doped and various N-doped MWNTs. Fig. 2d depicts weight loss of MWNTs as a function of temperature. According to the TGA results, it is well evident that nitrogen doping had a significant effect on the thermal stability of MWNTs. For instance, the inflection point was observed to be 650 °C for un-doped MWNTs, while N-doped MWNTs synthesized at 650 °C, 750 °C, and 850 °C exhibited inflection points equal to 517 °C, 590 °C and 578 °C, respectively.

It is well known that the defect formation and the presence of amorphous carbon result in lower oxidative thermal stability of graphitic structures. Therefore, the observed results are in good agreement with TEM and Raman spectra. It is worth noting that though the amount of defects presents within N-doped MWNTs synthesized at 750 °C and 850 °C were similar or less than un-doped MWNTs; however, their thermal stability was significantly poorer. Based on this comparison, one can conclude that the thermal stability is sensitive to various other parameters as well. This provides interesting outcomes where nitrogen doping facilitates oxidative degradation of MWNTs. Interestingly, the thermal stability was in line with the amount of nitrogen content in N-doped MWNTs. Similar observations have been reported where the presence of nitrogen resulted in inferior oxidative thermal stability.⁵⁷

The left-over residue following the tests corresponds to the alumina substrate and metal oxides.^58,59 In general, the synthesis yield is inversely proportional to the residue mass. Un-doped MWNTs depict smallest residue mass of 14.8%, manifesting highest yield. However, N-doped MWNTs resulted in lower yields. For instance, N-doped MWNTs synthesized at 850 °C showed residue mass of 21.8%. The catalyst particles comprised 80 wt% alumina and 20 wt% metallic particles. Alumina is insulative and metallic particles have much lower surface area than synthesized MWNTs. This implies that low yield can drastically deteriorate the final electrical properties of the nanocomposites. Given all the thermal parameters, one can conclude that the thermal stability and synthesis yield are mainly governed by the nitrogen content in N-doped MWNTs.

3.2. Dispersion of MWNTs in PVDF matrix

In case of polymer-based nanocomposites, the dispersion state of nanoparticles greatly controls the resultant properties of nanocomposites.⁶⁰ Therefore, effective dispersion of nanoparticles is an essential criterion to realize the exceptional properties associated with the functional nanoparticles. Owing to their nanoscopic size, MWNTs often agglomerate in polymer matrix. Therefore, the dispersion state of the nanoparticles in the PVDF matrix was assessed on microscopic level.

Fig. 3 shows the optical micrographs (OM) of the PVDF nanocomposites containing various MWNTs. The presence of micron sized agglomerates is well evident in all the PVDF nanocomposites. However, uniformly-dispersed, submicron-sized particles are present throughout the samples. These uniformly dispersed conducting nanoparticles can further lead to nomadic charge transference through tunneling or hopping mechanisms. Nevertheless, highly agglomerated nanoparticles lead to lower effective aspect ratio, therefore having adverse impact on electrical percolation threshold. Interestingly, the PVDF nanocomposites containing un-doped MWNTs showed better state of dispersion in the given matrix. However, poor dispersion was realized in the nanocomposites containing N-doped MWNTs. Similar observations have been reported earlier.^8,19 It is interesting to note that more agglomeration was realized in case of N-doped MWNTs synthesized at higher temperatures.

Fig. 3 Optical micrographs of PVDF nanocomposites containing (a) un-doped MWNTs; and N-doped MWNTs synthesized at (b) 650 °C, (c) 750 °C, and (d) 850 °C.

3.3. Electrical conductivity: connecting network of MWNTs

Since electrical conductivity and connecting network of MWNTs in host matrix play vital roles in microwave attenuation, room-temperature AC electrical conductivity was analyzed for the PVDF nanocomposites containing synthesized un-doped as well as N-doped MWNTs. In case of insulating materials, due to high resistance, there is no significant in-phase current flow. Therefore, at high frequencies, the governing mechanism for an observed conductivity is the dipole reorientation. In fact, with the increase of frequency and thus the decrease of time in each half cycle of alternating field, the current, i.e., charge over time, increases; therefore, AC conductivity shows an ascending trend with frequency. Whereas, in case of conductive materials, the in-phase current flow arising from nomadic charge dominates the dipole reorientation. This results in negligible alternate current electrical conductivity (σ_AC) over direct current electrical conductivity (σ_DC), manifesting frequency-independent conductivity.

In order to clearly understand the effect of N-doping and synthesis temperature, the concentration of MWNTs was fixed at 2.0 wt%. Fig. 4a shows the AC conductivity of the nanocomposites as a function of frequency. The nanocomposites containing un-doped MWNTs manifested highest electrical conductivity along with a frequency-independent plateau at lower frequencies, a characteristic of a semi-conductor. This is directly correlated to high aspect ratio and effective dispersion of un-doped MWNTs in the PVDF matrix.^27–29 Moreover, the highest synthesis yield in case of un-doped MWNTs also facilitates enhanced electrical conductivity in the PVDF nanocomposites. This phenomenon is strikingly different than what we observed for the conductivity of the powder samples, wherein the N-doped MWNTs (@850 °C) showed the highest bulk electrical conductivity.

Fig. 4 (a) AC electrical conductivity, (b) real permittivity, and (c) loss tangent as a function of frequency for PVDF nanocomposites.

In case of nanocomposites containing conductive nanoparticles, current flow is greatly governed by concentration of nanoparticles, which eventually decides the distance between discrete nanoparticles in a given matrix. Various charge transfer mechanisms are realized, namely conduction, hopping, and tunneling. For conduction, physical contact between nanoparticles is necessary, whereas for hopping and tunneling of charges physical contact is not necessary. Though insulating gap is present between conductive nanoparticles, sufficient field strength can allow charge to hop over or tunnel through the polymer gaps between nanoparticles. In this case, nomadic charge flow becomes comparable to dipole reorientation. Therefore, above a critical frequency, f_c, dipole reorientation is governing, whereas below f_c DC current dominates. Hence, in general, frequency-independent plateau can be observed especially at lower frequencies.⁹

The nitrogen doping of MWNTs resulted in strikingly decreased electrical conductivity of the nanocomposites at the given concentration of MWNTs. As observed earlier, nitrogen doping resulted in structural defects in MWNTs, thereby restricting transport of nomadic charges through the network of MWNTs. The resistance offered by defects and nitrogen results in disturbed pathway for nomadic charge transport. Moreover, relatively smaller aspect ratio of N-doped MWNTs hinders the formation of connecting network in the host matrix at any given concentration. As realized from structural analysis, N-doped MWNTs had a weaker crystalline structure; therefore, it can further lead to breakdown during melt-mixing process. In addition, poor dispersion state of nanoparticles in polymer matrix results in inferior connecting network.

Despite their high powder conductivity, nanocomposites containing 2.0 wt% N-doped MWNTs synthesized at 750 °C and 850 °C depicted insulating characteristic. It is interesting to note that though the intrinsic conductivity was strikingly higher for N-doped MWNTs (synthesized at 850 °C) than the un-doped MWNTs, the electrical conductivity of nanocomposites showed an inverse trend. Therefore, we can claim that the intrinsic conductivity of the MWNTs is governed by their crystalline structure, whereas the conductivity of the nanocomposites is controlled by MWNT physical properties and quality of interaction between MWNT and polymer matrix.

Another important parameter influencing charge storage and EMI shielding properties of a material is real part of permittivity. It represents charge storage ability of materials, and is greatly governed by charge polarization within the materials. Different polarization mechanisms, such as dipole and interfacial, contribute to charge storage and are strongly dependent on frequency.^61,62 Over the investigated frequency range, the dipole polarization arising from the PVDF matrix remained nearly constant, confirmed by dielectric spectroscopy of pure PVDF. Therefore, the resultant change in permittivity was mainly governed by the heterogeneous dispersion of highly conducting MWNTs in the PVDF matrix.

In general, nomadic charges get trapped at the interface between insulating PVDF and highly conducting MWNTs due to large mismatch between conductivities, and therefore, give rise to interfacial polarization. This phenomenon is mostly prominent at lower frequencies and disappears at higher frequencies due to small time span. Herein, due to large interfaces between PVDF and MWNTs, accumulation of nomadic charges gives rise to permittivity at lower frequencies. This effect, in general, is correlated to the formation of nano-capacitors within nanocomposites. In this context, highly conducting MWNTs act as conductors in an insulating polymer matrix, resulting in the formation of numerous nano-capacitors. This effect can be enhanced by increasing the concentration of MWNTs in the host matrix. Moreover, increase in MWNT content leads to smaller gap between them, enhancing polymer electronic polarization. Apart from interfacial and electronic polarization, MWNT structure itself results in charge storage and contributes towards high permittivity. Therefore, N-doped MWNTs are expected to have higher self-polarization than un-doped MWNTs. The self-polarization mechanism is prominent at higher frequencies where interfacial polarization effect is weak.

Fig. 4b depicts real permittivity as a function of frequency for various PVDF nanocomposites. The nanocomposites containing un-doped MWNTs showed highest real permittivity with a descending trend with frequency, characteristic of a nanocomposite close to or above the percolation threshold. The decline in real permittivity with frequency corresponds to interfacial relaxation effect at higher frequencies. Similar observations were revealed in nanocomposites containing N-doped MWNTs synthesized at the identical temperature, i.e. 650 °C. The PVDF nanocomposites containing N-doped MWNTs synthesized at higher temperatures, i.e. 750 °C and 850 °C, showed strikingly lower permittivity than that of un-doped MWNTs. This corresponds to the insulating nature of the nanocomposites where the percolation threshold was not achieved. In this case, the distance between individual MWNTs was significantly higher to generate nano-capacitors.

One can realize that the real permittivity and imaginary permittivity scale with electrical conductivity of the nanocomposites. Therefore, in case of nanocomposites containing un-doped MWNTs, i.e. semi-conducting nanocomposites, significantly higher permittivity and relaxation were recorded, whereas in case of insulating nanocomposites low real permittivity and no relaxation were observed. However, real permittivity was significantly higher than that of neat PVDF over the entire measured frequency range. This corresponds to the interfacial polarization and large self-polarization arising from N-doped MWNTs. The PVDF nanocomposites containing N-doped MWNTs synthesized at higher temperatures, i.e. 750 °C and 850 °C, manifested real permittivity of 23.15 and 18.2, respectively, at 10 Hz, whereas for neat PVDF real permittivity was only 11.4.

The retention of accumulated nomadic charges due to large polarization is a challenging task, and is represented by dissipation factor (tan δ = ε′′/ε′). In general, real permittivity (ε′) represents nomadic charge storage, whereas imaginary permittivity (ε′′) corresponds to the loss of nomadic charges. ε′′ manifests dissipation of energy and arises from ohmic loss.^63–65 In case of polymeric nanocomposites, ohmic loss arises from nomadic charge transfer through percolated network of MWNTs, whereas polarization loss is generated due to reorientation of electrical dipoles in each half cycle of AC field. As we discussed earlier, un-doped MWNTs and N-doped MWNTs synthesized at 650 °C formed percolating networks in the PVDF matrix at 2.0 wt%; therefore, the imaginary permittivity was mainly governed by ohmic loss and increased with decreasing frequency, characteristic of a semi-conductor or conductor.

In case of insulating PVDF nanocomposites, imaginary permittivity corresponds to nomadic charges in MWNTs. This is in good agreement with nanoparticle dispersion, where lack of connecting network limits ohmic losses. Therefore observed imaginary permittivity is mainly due to nomadic charges associated with MWNTs, which dissipate electrical energy. For charge storage applications, low dielectric loss is well appreciated,³⁹ whereas high dielectric losses is necessary for EMI shielding applications.^5,25 In this context, loss tangent (tan δ_ε) as a function of frequency was analyzed to understand the contributions of real and imaginary permittivities for various PVDF nanocomposites (Fig. 4c). Strikingly higher tan δ_ε of 129 was realized at 10 Hz in case of PVDF nanocomposites containing un-doped MWNTs, whereas the nanocomposites with N-doped MWNTs synthesized at 650 °C, 750 °C and 850 °C manifested tan δ_ε of 5.37, 0.04 and 0.02, respectively. One can easily correlate this with the network formation of MWNTs in given matrix. In case of nanocomposites containing N-doped MWNTs that were synthesized at higher temperature i.e. 750 °C and 850 °C, due to lack of network formation of MWNTs in given matrix, tan δ_ε was significantly smaller. In nanocomposites containing N-doped MWNTs that were synthesized at higher temperature i.e. 750 °C and 850 °C where network of MWNTs is not well developed the imaginary permittivity corresponds to the presence of nomadic charges in N-doped MWNTs. The latter nanocomposites manifested significantly enhanced real permittivity along with minimum losses over neat PVDF. This suggests both MWNTs polarization and interfacial polarization are effective and contribute toward higher charge storage ability. However, the nanocomposites containing N-doped MWNTs synthesized at lower temperature i.e. 650 °C, showed large losses. Therefore, owing to their enhanced real permittivity and minimum losses, the PVDF nanocomposites containing N-doped MWNTs synthesized at higher temperatures (i.e. 750 °C and 850 °C) can be used as efficient charge storage materials.

These results indicate that by the help of self-polarizing sites generated by nitrogen incorporation in MWNTs structure low dissipation factor can be achieved. Since high real permittivity and low imaginary permittivity, i.e. small tan δ_ε, are essential criteria for charge storage application, the nanocomposites containing N-doped MWNTs depict promising results as potential candidates for charge storage. It is important to note that the intrinsic conductivities of MWNTs synthesized at higher temperatures (i.e. 750 °C and 850 °C) was significantly higher than that of un-doped MWNTs, resulting large driving force for charge accumulation at polymer/MWNTs interface. The accumulated charge can be effectively retained by avoiding electrical percolation in the host matrix. Hence, the insulating nanocomposites with high storage ability can be designed by taking advantage of self-polarizable N-doped MWNTs.

3.4. Attenuation of microwave radiation: effect of nitrogen doping and synthesis temperature

Recently, there is an immense surge in materials research to design EMI shielding materials using flexible, tailor-made polymer nanocomposites. Over the years, it has been well understood that the microwave attenuation is mainly governed by the dielectric properties of developed nanocomposites, and this can be controlled by intrinsic dielectric properties of matrix and used fillers.^4,66,67 According to the microwave theory, the microwave attenuation takes place due to large dielectric losses. Various attempts have been carried out to manipulate the dielectric properties of polymer nanocomposites, where dispersion of high-dielectric-constant ceramic particles^68–70 and conducting inclusions^71–74 manifested promising results in microwave attenuation. Since electrical conductivity is an essential property for microwave attenuation, carbonaceous conducting nanoparticles have received great attention. In this context, well-dispersed, high-aspect-ratio conducting inclusions like MWNTs in host polymer matrix have shown impressive results in microwave attenuation. However, we still believe that great efforts are required to design multifunctional nanocomposites for high-performance microwave shielding.

In case of polymer nanocomposites containing MWNTs, at the percolation threshold, microwave attenuation can be strikingly enhanced by connecting network of MWNTs (ohmic loss). In fact, at enhanced conductive network, the nomadic charges have more mean free paths to go through, and therefore, can dissipate more electrical energy.⁷⁵ Similarly, localized polarization centers generated by MWNTs, due to their distorted structure, might result in further attenuation by multiple scattering. However, in case of N-doped MWNTs, though the distorted structure leads to localized polarization, it also results in low intrinsic conductivity of MWNTs by blocking pathway of nomadic charge transport. Enhanced microwave attenuation can be achieved by large ohmic losses through nomadic charge transfer and polarization losses deriving from distorted structure of MWNTs. Therefore, the balance between ohmic losses (i.e. electrical conductivity and conductive network formation) and polarization losses (i.e. distorted structure of MWNTs due to presence of nitrogen) could be an ideal recipe to achieve enhanced microwave attenuation.

Herein, our main concern is to optimize the conductive network to achieve large ohmic losses and localized polarization losses to achieve synergistic effect for microwave attenuation. To understand the effect of the dielectric properties of MWNTs on the microwave attenuation properties, MWNTs were tailored by doping with nitrogen. As discussed earlier, dielectric properties were systematically tuned by varying the synthesis temperature. To the best of our knowledge, no work has concentrated on the systematic study of dielectric and microwave attenuation properties of N-doped MWNTs in this frame of reference, i.e. varying the synthesis temperature.

Since attenuation of microwave radiation is mainly governed by dielectric properties, the complex permittivity was assessed in X (8.2 GHz to 12.4 GHz) and K_u-band (12.4 GHz to 18 GHz) in the PVDF nanocomposites. Fig. 5a shows real and imaginary permittivity of PVDF nanocomposites containing un-doped and various N-doped MWNTs. Interestingly, nanocomposites depicted similar trends of real and imaginary permittivity in GHz frequencies as realized in low frequencies (<1 MHz), where nanocomposites consisting of un-doped MWNTs manifested highest real permittivity along with large dielectric losses. The effect of nitrogen doping was well evident in the complex permittivity data even at high frequencies. The real and imaginary parts of permittivity significantly decreased with nitrogen doping. Moreover, it was observed that the effect was more prominent for N-doped MWNTs synthesized at higher synthesis temperatures. Since dielectric loss greatly influences the microwave attenuation, the large dielectric loss realized in nanocomposites containing un-doped MWNTs can facilitate the microwave attenuation.

Fig. 5 (a) Complex permittivity (solid symbols correspond to real permittivity and open symbols represent imaginary permittivity); (b) SE as a function of frequency for PVDF nanocomposites; (c) SE as a function of thickness for PVDF nanocomposites.

The total shielding effectiveness (SE) is estimated by scattering parameters using the following relationship.⁵

where scattering parameter (S₁₂) represents reverse transmission coefficient. (S₁₂)² corresponds to the ratio of power transmitted through shield to incident power.

Fig. 5b depicts SE for PVDF nanocomposites containing 2.0 wt% un-doped and N-doped MWNTs. The nanocomposites containing un-doped MWNTs showed highest SE of −7.77 dB, manifesting more than 80% attenuation of incoming microwave radiation at 18 GHz. The nanocomposites containing N-doped MWNTs resulted in lower SE. The SE results were in good agreement with dielectric losses of the nanocomposites. As discussed earlier, nitrogen doping severely affects structural integrity of MWNTs and also largely decreases the aspect ratio of N-doped MWNTs. This results in lower electrical conductivity of nanocomposites. It was also realized that the SE was in line with electrical conductivity of the nanocomposites rather than that of intrinsic conductivity of synthesized MWNTs.

Moreover, one can easily distinguish the effect of synthesis temperature used for the N-doped MWNTs on the SE. The SE was decreased with increasing synthesis temperature for N-doped MWNTs. For instance, PVDF nanocomposites containing N-doped MWNTs synthesized at higher temperature, i.e. 850 °C, manifested SE of −3.38 dB at 18 GHz. However, PVDF nanocomposites containing N-doped MWNTs that were synthesized at 650 °C manifested SE of −4.33 dB at 18 GHz. The higher SE for the latter nanocomposites corresponds to the higher electrical conductivity of nanocomposites in addition to local polarizing centers generated due to doping.

In order to understand the effect of shield thickness on the microwave attenuation properties, EMI shielding experiments were performed on specimens with different thicknesses and scattering parameters were recorded using the VNA. Fig. 5c shows SE as a function of thickness for various PVDF nanocomposites at 18 GHz. It was realized that SE scaled with shield thickness. Similar outcomes have been reported in the literature.^11,76,77 The nanocomposites containing N-doped MWNTs manifested smaller shielding effectiveness than those of un-doped MWNTs at any given thickness. The rate of increase of SE with increasing thickness was strikingly higher for nanocomposites containing un-doped MWNTs. These observations reveal that the electrical conductivity of the nanocomposites governs the total microwave attenuation.

In order to realize the effect of MWNTs concentration on SE, PVDF nanocomposites containing 3.0 wt% un-doped as well as N-doped MWNTs were prepared and SE was assessed. Table 2 shows SE for PVDF nanocomposites containing 2.0 wt% and 3.0 wt% un-doped and N-doped MWNTs at reference frequency of 18 GHz. As expected, SE was higher at higher concentration of MWNTs. This mainly corresponds to the enhanced nomadic charge transfer through dense MWNTs network at higher concentration, resulting in large dielectric losses arising from nomadic charge transfer and polarization.

Table 2 The SE for PVDF nanocomposites containing un-doped and N-doped MWNTs at 18 GHz for 5 mm thick specimen

PVDF composites containing	Concentration (wt%)	SE (dB)
Un-doped (at 650 °C)	2.0	−7.77
N-Doped (at 650 °C)	2.0	−4.33
N-Doped (at 750 °C)	2.0	−4.48
N-Doped (at 850 °C)	2.0	−3.38
Un-doped (at 650 °C)	3.0	−12.2
N-Doped (at 650 °C)	3.0	−6.3
N-Doped (at 750 °C)	3.0	−7.5
N-Doped (at 850 °C)	3.0	−5.8

---

Taking into account all the parameters, we can conclude that the nanocomposites containing N-doped MWNTs (synthesized at lower temperatures) manifest moderate charge storage ability, electrical conductivity and microwave attenuation performance. Interestingly, synthesis of N-doped MWNTs at higher temperature, i.e. 850 °C, leads to highly crystalline, less-defective MWNTs with low aspect ratio. This results in strikingly higher charge storage; however, electrical conductivity and microwave attenuation decreases extensively. Therefore, synthesis temperature plays a vital role in deciding intrinsic properties of MWNTs as well as properties of nanocomposites. Hence, with suitable choice of synthesis temperature, we can effectively tune the charge storage and microwave attenuation properties of nanocomposites. In conclusion, this study shows that nitrogen doping and synthesis temperature can be used as regulative tools to manipulate charge storage and microwave attenuation performance of MWNT nanocomposites.

4. Conclusions

Flexible, lightweight PVDF nanocomposites containing self-polarizable MWNTs were developed for charge storage and EMI shielding applications. Our study showed that nitrogen doping, as a unique strategy, can pave way for developing highly defective structures, resulting in numerous polarizing centers. The N-doped MWNTs synthesized at different temperatures resulted in bamboo-like structure, whereas un-doped MWNTs had an open-channel microstructure. Furthermore, it was revealed that the electrical properties of synthesized MWNTs are mainly governed by structural defects, whereas electrical properties of nanocomposites are controlled by the aspect ratio and heterogeneous dispersion of MWNTs in the host matrix. In this study, we have systematically tailored the polarizing centers present on the MWNTs by controlling the synthesis temperature. The nanocomposites containing N-doped MWNTs depicted significantly smaller dissipation factor over un-doped MWNTs, manifesting excellent charge storage ability. The nanocomposites containing un-doped MWNTs provided higher microwave attenuation due to improved electrical conductivity of the nanocomposites. Moreover, EMI SE was decreased with increasing synthesis temperature for N-doped MWNTs. It was also revealed that the microwave attenuation was governed much more by electrical conductivity of the nanocomposites rather than intrinsic conductivity of MWNTs. From this study, it was understood that the fine balance between electrical conductivity of the nanocomposites and self-polarization due to doping governs important properties like charge storage ability and microwave attenuation in polymer nanocomposites. Hence, with proper selection of synthesis temperature for N-doped MWNTs effective charge storage and EMI shielding materials can be developed.

Acknowledgements

Department of Science and Technology is gratefully acknowledged for providing the financial aid. We thank the Natural Sciences and Engineering Research Council of Canada (NSERC) for funding. We thank Drs Petra Pötschke and Beate Krause from Leibniz Institute of Polymer Research Dresden (IPF), Germany for their great assistance with powder conductivity measurement.

References

1. F. Hussain, M. Hojjati, M. Okamoto and R. E. Gorga, J. Compos. Mater., 2006, 40, 1511–1575.
2. J. Jordan, K. I. Jacob, R. Tannenbaum, M. A. Sharaf and I. Jasiuk, Mater. Sci. Eng., A, 2005, 393, 1–11.
3. M. Moniruzzaman and K. I. Winey, Macromolecules, 2006, 39, 5194–5205.
4. J.-M. Thomassin, C. Jérôme, T. Pardoen, C. Bailly, I. Huynen and C. Detrembleur, Mater. Sci. Eng., R, 2013, 74, 211–232.
5. M. H. Al-Saleh, W. H. Saadeh and U. Sundararaj, Carbon, 2013, 60, 146–156.
6. M. H. Al-Saleh and U. Sundararaj, Carbon, 2009, 47, 1738–1746.
7. D. Chung, J. Mater. Eng. Perform., 2000, 9, 350–354.
8. M. Arjmand, K. Chizari, B. Krause, P. Pötschke and U. Sundararaj, Carbon, 2016, 98, 358–372.
9. W. Bauhofer and J. Z. Kovacs, Compos. Sci. Technol., 2009, 69, 1486–1498.
10. H. Kim, K. Kim, C. Lee, J. Joo, S. Cho, H. Yoon, D. Pejaković, J.-W. Yoo and A. Epstein, Appl. Phys. Lett., 2004, 84, 589–591.
11. S. P. Pawar, S. Kumar, A. Misra, S. Deshmukh, K. Chatterjee and S. Bose, RSC Adv., 2015, 5, 17716–17725.
12. S. P. Pawar, K. Pattabhi and S. Bose, RSC Adv., 2014, 4, 18842–18852.
13. H. Zhang, W. Ding and D. K. Aidun, J. Nanosci. Nanotechnol., 2015, 15, 1660–1668.
14. E. Zussman, X. Chen, W. Ding, L. Calabri, D. Dikin, J. Quintana and R. Ruoff, Carbon, 2005, 43, 2175–2185.
15. A. Saha, C. Jiang and A. A. Martí, Carbon, 2014, 79, 1–18.
16. C. Jiang, A. Saha and A. Martí, Nanoscale, 2015, 7, 15037–15045.
17. D. A. Nguyen, Y. R. Lee, A. V. Raghu, H. M. Jeong, C. M. Shin and B. K. Kim, Polym. Int., 2009, 58, 412–417.
18. Y. Breton, G. Desarmot, J. Salvetat, S. Delpeux, C. Sinturel, F. Beguin and S. Bonnamy, Carbon, 2004, 42, 1027–1030.
19. M. Arjmand and U. Sundararaj, Compos. Sci. Technol., 2015, 118, 257–263.
20. S. P. Pawar, V. Bhingardive, A. Jadhav and S. Bose, RSC Adv., 2015, 5, 89461–89471.
21. M. Sharma, S. Sharma, J. Abraham, S. Thomas, G. Madras and S. Bose, Mater. Res. Express, 2014, 1, 035003.
22. T. K. Gupta, B. P. Singh, S. R. Dhakate, V. N. Singh and R. B. Mathur, J. Mater. Chem. A, 2013, 1, 9138–9149.
23. J. Yun, J. S. Im, Y.-S. Lee and H.-I. Kim, Eur. Polym. J., 2010, 46, 900–909.
24. J. S. Im, I. J. Park, S. J. In, T. Kim and Y.-S. Lee, J. Fluorine Chem., 2009, 130, 1111–1116.
25. Y. Huang, N. Li, Y. Ma, F. Du, F. Li, X. He, X. Lin, H. Gao and Y. Chen, Carbon, 2007, 45, 1614–1621.
26. J. Li, P. C. Ma, W. S. Chow, C. K. To, B. Z. Tang and J. K. Kim, Adv. Funct. Mater., 2007, 17, 3207–3215.
27. S.-H. Yao, Z.-M. Dang, M.-J. Jiang, H.-P. Xu and J. Bai, Appl. Phys. Lett., 2007, 91, 212901.
28. D. Wu, L. Wu, W. Zhou, Y. Sun and M. Zhang, J. Polym. Sci., Part B: Polym. Phys., 2010, 48, 479–489.
29. S. Bose, A. R. Bhattacharyya, A. P. Bondre, A. R. Kulkarni and P. Pötschke, J. Polym. Sci., Part B: Polym. Phys., 2008, 46, 1619–1631.
30. K. Chizari, A. Vena, L. Laurentius and U. Sundararaj, Carbon, 2014, 68, 369–379.
31. K. Chizari and U. Sundararaj, Mater. Lett., 2014, 116, 289–292.
32. J. W. Jang, C. E. Lee, S. C. Lyu, T. J. Lee and C. J. Lee, Appl. Phys. Lett., 2004, 84, 2877–2879.
33. S. Ci, Z. Wen, J. Chen and Z. He, Electrochem. Commun., 2012, 14, 71–74.
34. Y. T. Lee, N. S. Kim, S. Y. Bae, J. Park, S.-C. Yu, H. Ryu and H. J. Lee, J. Phys. Chem. B, 2003, 107, 12958–12963.
35. K. Chizari, I. Janowska, M. Houllé, I. Florea, O. Ersen, T. Romero, P. Bernhardt, M. J. Ledoux and C. Pham-Huu, Appl. Catal., A, 2010, 380, 72–80.
36. S. Trasobares, O. Stephan, C. Colliex, W. Hsu, H. Kroto and D. Walton, J. Chem. Phys., 2002, 116, 8966–8972.
37. S. Van Dommele, A. Romero-Izquirdo, R. Brydson, K. De Jong and J. Bitter, Carbon, 2008, 46, 138–148.
38. H. C. Choi, J. Park and B. Kim, J. Phys. Chem. B, 2005, 109, 4333–4340.
39. M. Arjmand and U. Sundararaj, ACS Appl. Mater. Interfaces, 2015, 7, 17844–17850.
40. Y. T. Lee, N. S. Kim, J. Park, J. B. Han, Y. S. Choi, H. Ryu and H. J. Lee, Chem. Phys. Lett., 2003, 372, 853–859.
41. M. Glerup, M. Castignolles, M. Holzinger, G. Hug, A. Loiseau and P. Bernier, Chem. Commun., 2003, 2542–2543.
42. F. Ding, A. Rosén and K. Bolton, J. Chem. Phys., 2004, 121, 2775–2779.
43. Y. Kobayashi, H. Nakashima, D. Takagi and Y. Homma, Thin Solid Films, 2004, 464, 286–289.
44. A. F. Kibria, Y. Mo, K. Nahm and M. Kim, Carbon, 2002, 40, 1241–1247.
45. N. Alexeyeva, E. Shulga, V. Kisand, I. Kink and K. Tammeveski, J. Electroanal. Chem., 2010, 648, 169–175.
46. S. Kundu, T. C. Nagaiah, W. Xia, Y. Wang, S. V. Dommele, J. H. Bitter, M. Santa, G. Grundmeier, M. Bron and W. Schuhmann, J. Phys. Chem. C, 2009, 113, 14302–14310.
47. S. Kundu, W. Xia, W. Busser, M. Becker, D. A. Schmidt, M. Havenith and M. Muhler, Phys. Chem. Chem. Phys., 2010, 12, 4351–4359.
48. K. Ghosh, M. Kumar, T. Maruyama and Y. Ando, J. Mater. Chem., 2010, 20, 4128–4134.
49. S. Maldonado, S. Morin and K. J. Stevenson, Carbon, 2006, 44, 1429–1437.
50. H. R. Barzegar, E. Gracia-Espino, T. Sharifi, F. Nitze and T. Wågberg, J. Phys. Chem. C, 2013, 117, 25805–25816.
51. M. S. Dresselhaus, G. Dresselhaus, R. Saito and A. Jorio, Phys. Rep., 2005, 409, 47–99.
52. S. Osswald, E. Flahaut, H. Ye and Y. Gogotsi, Chem. Phys. Lett., 2005, 402, 422–427.
53. S. Brown, A. Jorio, M. Dresselhaus and G. Dresselhaus, Phys. Rev. B: Condens. Matter Mater. Phys., 2001, 64, 073403.
54. L. Bulusheva, A. Okotrub, I. Kinloch, I. Asanov, A. Kurenya, A. Kudashov, X. Chen and H. Song, Phys. Status Solidi B, 2008, 245, 1971–1974.
55. J. D. Wiggins-Camacho and K. J. Stevenson, J. Phys. Chem. C, 2009, 113, 19082–19090.
56. T. Sharifi, F. Nitze, H. R. Barzegar, C.-W. Tai, M. Mazurkiewicz, A. Malolepszy, L. Stobinski and T. Wågberg, Carbon, 2012, 50, 3535–3541.
57. E. N. Nxumalo, P. J. Letsoalo, L. M. Cele and N. J. Coville, J. Organomet. Chem., 2010, 695, 2596–2602.
58. E. Antunes, V. De Resende, U. Mengui, J. Cunha, E. Corat and M. Massi, Appl. Surf. Sci., 2011, 257, 8038–8043.
59. E. Edwards, E. Antunes, E. C. Botelho, M. Baldan and E. Corat, Appl. Surf. Sci., 2011, 258, 641–648.
60. S. Pegel, P. Pötschke, G. Petzold, I. Alig, S. M. Dudkin and D. Lellinger, Polymer, 2008, 49, 974–984.
61. M. Antunes, M. Mudarra and J. I. Velasco, Carbon, 2011, 49, 708–717.
62. J.-K. Yuan, S.-H. Yao, Z.-M. Dang, A. Sylvestre, M. Genestoux and J. Bai, J. Phys. Chem. C, 2011, 115, 5515–5521.
63. S. H. Foulger, J. Appl. Polym. Sci., 1999, 72, 1573–1582.
64. F. He, S. Lau, H. L. Chan and J. Fan, Adv. Mater., 2009, 21, 710–715.
65. P. C. Watts, W. K. Hsu, A. Barnes and B. Chambers, Adv. Mater., 2003, 15, 600–603.
66. S. P. Pawar, S. Biswas, G. P. Kar and S. Bose, Polymer, 2016, 84, 398–419.
67. M.-S. Cao, X.-X. Wang, W.-Q. Cao and J. Yuan, J. Mater. Chem. C, 2015, 3, 6589–6599.
68. Z. M. Dang, Y. H. Lin and C. W. Nan, Adv. Mater., 2003, 15, 1625–1629.
69. Y. Bai, Z.-Y. Cheng, V. Bharti, H. Xu and Q. Zhang, Appl. Phys. Lett., 2000, 76, 3804–3806.
70. M. Arbatti, X. Shan and Z. Y. Cheng, Adv. Mater., 2007, 19, 1369–1372.
71. Z. Spitalsky, D. Tasis, K. Papagelis and C. Galiotis, Prog. Polym. Sci., 2010, 35, 357–401.
72. R. Khare and S. Bose, J. Miner. Mater. Charact. Eng., 2005, 4, 31–46.
73. S. Stankovich, D. A. Dikin, G. H. Dommett, K. M. Kohlhaas, E. J. Zimney, E. A. Stach, R. D. Piner, S. T. Nguyen and R. S. Ruoff, Nature, 2006, 442, 282–286.
74. M. H. Al-Saleh and U. Sundararaj, Carbon, 2009, 47, 2–22.
75. M. Arjmand, T. Apperley, M. Okoniewski and U. Sundararaj, Carbon, 2012, 50, 5126–5134.
76. S. H. Kim, S. H. Jang, S. W. Byun, J. Y. Lee, J. S. Joo, S. H. Jeong and M. J. Park, J. Appl. Polym. Sci., 2003, 87, 1969–1974.
77. P. Jana and S. De, IEEE Trans. Electromagn. Compat., 1992, 34, 478–481.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra15037c

‡ SPP and MA made equal contribution to this work.

---