Improved Thermal and Electrical Properties of Nanocomposites Based on Poly(vinyl pyrrolidone)/Poly(acrylonitrile)/Multiwalled Carbon Nanotubes

Abstract

Poly(vinyl pyrrolidone)/Poly(acrylonitrile)/multi-walled carbon nanotubes composites (PVP/PAN/MWNTs) were fabricated by a simple solution cast technique with a wide composition range from 0 to 3.12 wt% of MWNTs. The nanocomposites were characterized by fourier transform infrared (FT-IR) spectroscopy and scanning electron microscopy (SEM). A percolated MWNTs network structure and relatively good dispersion of MWNTs are evident in PVP/PAN/MWNTs nanocomposites with 3.12 wt% of MWNTs. The thermal properties of the polymer blend with MWNTs were carried out by means of different scanning calorimetry (DSC). It indicated that the incorporation of MWNTs had a significant influence on crystallization and melting behaviors for the polymer nanocomposites. The four point probe method was used to measure the electrical conductivity, and the result showed ionic conductivity of the order of 2.87 × 10⁻⁴ S/cm to 1.91 × 10⁻² S/cm. The effect of the concentration of the filler on the conductivity of the polymer nanocomposite was discussed. Nanocomposites based on PVP/PAN and MWNTs as filler show a significant enhancement in the electrical conductivity as a function of temperature. The low percolation threshold in PVP/PAN/MWNTs nanocomposites was explained by high aspect ratio of the nanotubes and their distribution inside the polymer blend.

1. Introduction

Carbon nanotubes (CNTs), first prepared by Minima in 1991 [1], have attracted considerable interest owing to their exceptional electrical, physical and mechanical properties [2–6]. These nano-structural elements can also be used as nano-fillers and nanoreinforcements of advanced composite materials to improve the mechanical, thermal and impact-resistance properties. CNTs can be considered as attractive candidates in diverse nanotechnological applications, such as fillers in polymer matrixes, molecular tanks, biosensors, and many others. However, the lack of solubility and the difficult manipulation in any solvents have imposed great limitations to the use of CNTs. In fact, it is very hard for CNTs to dissolve completely in organic solvents and aqueous solutions. Although they can be dispersed in some solvents by sonication, precipitation always immediately occurs when this process is interrupted. On the other hand, it has been demonstrated that CNTs can interact with different classes of compounds [7].

It has been paid great attention that incorporation of nanoparticles into a polymer matrix to form composite materials where nanoparticles are randomly dispersed [8–11]. In this regard, polymers are a good choice as host materials, since they have a wide variety of bulk physical properties and possess flexible processability [12]. There have been several works about the preparation of new nanoparticles/polymer composites until now [13–17]. Solution-based casting methods provide an advantage through low viscosities, which could facilitate mixing and dispersion of the CNTs. Many studies used these methods to process both thermosetting and thermoplastic polymers. Shaffer et al. [18] firstly dispersed chemically oxidized catalytic MWCNTs in water. And then the dispersed MWCNTs were carefully blended with solutions of polyvinyl alcohol in water to give composite dispersions which could be drop-cast to form films with up to 60 wt% nanotubes. Jin et al. [19] produced various types of polymer-coated and polymer-grafted MWNTs solutions, in some cases evaporating the solvent and subsequently melt-mixing with another polymer. Yudasaka et al. [20] used a mixture of single wall nanotubes (SWNTs) and PMMA in monochlorobenzene (MCB) for dispersion, purification and subsequent spin-casting of the material. CNTs often play important role in their nanocomposites due to their excellent electrical and thermal properties. For example, the composites of CNTs and polymer could have good conductivity with extremely low CNTs contents. These composites exhibit percolation-type behavior. A wide range of values has been reported for percolation thresholds and conductivity of CNTs composites, depending on the processing method, polymer matrix and nanotube type [21]. Electrical conductivity can be explained by the established percolation theory [22], with an onset of the conductivity when a critical filler concentration, commonly named percolation threshold, is reached to form conductive pathways. The percolation threshold of nanotubes filled composites primarily is dependent on the processing methods, eventually on the dispersion of the SWNTs or MWNTs in a matrix. Direct blend of carbon nanotubes and a thermoplastic polymer by melt processing results in a high threshold since the dispersion of the very thin tubes is not facile. Meincke et al. [23] reported that a binary composite of MWNTs and polyamide-6 prepared by melt processing has a percolation of -6 wt% with a conductivity of 10⁻⁷ S/cm at the loading level. Maitra et al. [24] reported the electrical conductivity of PAN/MWNTs. He observed that the conductivity increased almost two orders of magnitude upon adding 0.5 wt% CNTs. Akhtar et al. [25] concluded that the PAN/CNT composite electrolyte showed high amorphicity and ionic conductivity due to the uniform dispersion and strong bonding of PAN on the surface of CNTs via thermal polymerization. Regev et al. [26] prepared SWNT/SDS/PS composites using latex technology, which exhibits an excellent conductive property with a percolation threshold of 0.28 wt%.

Miscible polymer blends are attracting host materials to which CNTs can be inserted because this type of mixtures has a degree of mixing down to the molecular level. Many literature reports deal with the addition of CNTs as fillers into polymer matrices; however, just a few investigate the effect of adding CNTs to a polymer matrix specifically formed by a miscible polymer blend. In this work, we investigated the effects of MWNTs with different ratios on the morphology, thermal and electrical properties of PVP/PAN blends. The influence of temperature and mass percent of MWNTs on the PVP/PAN blends were also discussed.

2. Materials and Methods

2.1 Materials

Poly(vinyl pyrrolidone) (PVP) with an average molecular weight of 30 × 10³ was obtained from Fluka, poly(acrylonitrile) (PAN) of molecular weight 9.5 × 10⁴ and dimethylformamide (DMF) form Aldrich Chemical Co. Inc. (Milwaukee, Wisconsin), while multi-walled carbon nanotubes (MWNTs) were supplied by Nanocyl S.A (Hamburg, Germany) with diameter 10 nm, length 0.1–10 micron, a Brunauer-Emmett-Teller (BET) surface area of 250–300 m²/g, and metal oxide (impurity) content of about 10% and purity 90%.

The blends were prepared by solution casting using DMF as solvent. In the first part of the work, PVP and PAN were blended in several weight percent ratios and dissolved in DMF. The solutions were stirred overnight and then poured into glass dishes and allowed to evaporate slowly at room temperature. The solid films were further dried under vacuum to remove residual solvent. The selected weight ratio PVP:PAN (50:50) was selected for the second part of the experiments. Here, different weights of MWNTs (0, 0.71, 1.31, 1.90 and 3.12 wt%) were mixed into the chosen ratio of PVP:PAN.

2.2 Methods

For the characterization of the samples, a Fourier transforms infrared spectrometer (FTIR) was carried on Nicolet DX 510 FTIR, using KBr pellet (AR grade) at room temperature. Differential Scanning Calorimetry (DSC) measurements were carried out on a Dupont 2000 DSC in standard mode with nitrogen gas. All samples were weighed in the range of 5 to 10 mg and heated to 350 °C using a heating rate of 10 °C/min. The morphology of the composites was observed with SEM (model JSM-6400, JEOL (Peabody, MA); low voltage of 20 kV). Scanning electron microscopy (SEM) samples were sputtered by 20nm gold before observation. The electrical conductivity of the samples was measured at different temperatures by using a four-probe technique (EPP 0602, Entek electronics).

3. Results and Discussions

The effect of nanoparticles on morphology and properties of polymer blends has attracted great interest because of the improved physical properties as compared with unmodified polymers. The SEM images of the PVP/PAN composite with 3.12%wt MWNTs are given in Fig. 1 (a–d). Fig. 1a shows that pure MWNTs are very long and highly entangled in the solid state to form a dense, robust, network structure. The diameters of the MWNTs are in the range of 10–15 nm.

Fig. 1.

SEM micrographs for the (a) pure MWNTs and (b, c and d) PVP/PAN/MWNTs in different scales.

The Fig. 1b, 1c and 1d show quite uniform spherical structure of PVP/PAN/MWNTs. There is no obvious aggregation of MWNTs in this sample. After adding PVP we could find that PVP significantly improved the dispersion of MWNTs and encouraged denser MWNTs network structure in PAN matrix through the special interaction between MWNTs and PVP and the hydrogen bonding effect between PAN and PVP. During the sonication process, PVP/PAN may wrap these tubes to prevent the aggregation of MWNTs upon the evaporation DMF. Moreover, the affinity of PVP to MWNTs plays an important role in good dispersion of MWNTs in PVP/PAN/DMF solution. It is noted that MWNTs are quite easily dispersed in PVP/PAN/DMF solutions. PVP is quite well soluble in water as well as in other solvents, so it is one of good polymer additives as a pore forming agent and could greatly improve the dispersion of MWNTs in polymer matrix [27]. It has been known that PVP is well miscible with PAN. According to Guo et al. [28], the blends of PAN/PVP are compatible over all compositions. In fact, it has been widely studied that the effect of PVP addition as a pore-forming agent to the casting polymer solutions on the formation of phase inversion membrane [29].

FTIR spectra of PVP, PAN, PVP/PAN and PVP/PAN/MWNTs nanocomposites are shown in Fig. 2. Several characteristic IR vibrational bands at 1,668 cm⁻¹ (C=O stretching), 1,373 cm⁻¹ (CH₂ bending and deformation), and 1,290 cm⁻¹ (C-N stretching) confirm organic groups of PVP. Characteristic peak at 2,214 cm⁻¹ is due to the stretching vibration of cyano group (CN), 1,454 cm⁻¹ for (-CH₃) and 1,373 cm⁻¹ (CH₂), which can be observed in PAN. The PVP/PAN/MWNTs spectrum shows absorbance peaks at 1,449 cm⁻¹ (-CH₃), 2,243 cm⁻¹ (CN), 1,373 cm⁻¹ (-CH₂-) and 3489 cm⁻¹ (O–H) characteristic of polymer composites. The shift in the carbonyl peak from 1668 to 1,670 cm⁻¹ in PVP is attributed to the formation of the intermolecular bonds in the PVP/PAN blend.

Fig. 2.

FTIR spectra of (a) PVP, (b) PAN, (c) PVP/PAN and (d) PVP/PAN/MWNTs nanocomposites.

The inset of Fig. 2 demonstrates a small shift of the C≡N and OH peak following MWNTs embedding compared with the pure PAN and PVP/PAN blend. The C≡N band, originally appearing at 2,214 cm⁻¹, is slightly shifted to 2,243 cm⁻¹ for PVP/PAN/MWNTs nanocomposites, however it is evident that the OH peak for the PVP/PAN/MWNTs (3,489 cm⁻¹) is shifted compared to the PVP/PAN blend (3,448 cm⁻¹). DSC analysis is generally one of the most convenient methods to analyze thermal behaviour of polymers and the composites. Fig. 3 shows typical DSC curves of PVP/PAN and PVP/PAN/MWNTs nanocomposite at the temperature range from 25–325 °C. PVP shows an endothermic peak between 82 and 150 °C (peak temperature), which is similar to the value reported in the literature [30].

Fig. 3.

DSC curves for (a) PVP/PAN blend and (b) PVP/PAN/MWNT nanocomposites containing 3.12 wt% MWNTs.

However PAN exhibits a sharp exothermic peak between about 285 °C and 300 °C. This peak corresponds to dehydrogenation, oxidation, cross-linking and mainly cyclization reactions. The effect of PVP/PAN blend composition on the melting temperature of PVP/PAN is given in Table 1. It can be seen that there is an apparent decrease in the melting temperature of the PVP/PAN with the increase of PVP content in PVP/PAN film from 295.81 °C to 276.63 °C, which means the decrease of PAN crystallinity in PVP/PAN blends. This phenomenon may be explicable in terms of thermodynamic mixing accompanied by exothermic interaction between a crystalline polymer and an amorphous polymer [31]. It is a significance of kinetic hindrance enforced by amorphous polymer upon crystallization process. Obviously, the kinetic effects can be affected by conventional dilution of crystalline polymer with amorphous component and the existence of specific interaction between polymers improves these effects [31]. This effect of amorphous fractions on crystalline regions was also observed in thermal properties of PEO/PVP blend [32], where the size of crystallinity peaks smaller and T_m was lower than that of pure PEO.

Table 1.

Thermal properties of PVP/PAN/MWNTs nanocomposites.

Sample	ΔHm (J/g)	Tm (°C)	Tc (°C)
PVP		117.62
PAN	343.6	295.18
PVP/PAN (30:70 wt%)	229.6	276.63
PVP/PAN (50:50 wt%)		279.46	65.0
PVP/PAN (70:30 wt%)	187.4	291.39	103.20
PVP/PAN/MWNTs (0.71 wt% of MWNTs)	199.3	291.79	80.0
PVP/PAN/MWNTs (1.51 wt% of MWNTs)	142.7	293.25	89.65
PVP/PAN/MWNTs (1.90 wt% of MWNTs)		300.30	70.11
PVP/PAN/MWNTs (3.12 wt% of MWNTs)		301.56	103.35

Ongoing from the polymer blends to the polymer nanocomposites with MWNTs, it has been evidently observed that incorporation of MWNTs has a significant influence on crystallization and melting behaviors for the polymer composites. The crystallization temperature (T_c) and melting temperature (T_m) of PVP/PAN blend (50 wt% PVP and 50 wt% PAN) as a function of MWNTs concentration are represented in Fig. 4. As seen, incorporation of MWNTs not only results in the increase of melting and decrease of crystallization temperatures but also reduces the crystallinity of PAN.

Fig. 4.

(a) melting and (b) crystallization temperature of PVP/PAN/MWNT nanocomposites as a function of MWNT content.

It can be concluded that incorporation of MWCNT caused a slow growth in crystals. This effect may be explained by the fact that adding carbon nanotubes to the blend inhibits crystallization. It was observed that, the melting temperature of the PVP/PAN film increased from 279.46 °C to approximately 301 °C in the blend with increasing of the MWNTs content (0–3.12 wt%). Similar observation was obtained by Nam et al [33]. It was reported that the T_m of PVDF blended by MWNTs at various contents (0.01–5 wt%) increased from 177.9 °C to 182.1 °C.

Measurements of electrical properties of PVP/PAN/MWNTs nanocomposites have been carried out at different concentration of MWNTs in temperature range 293–353 K by using four probe technique. Electrical conductivity was evaluated using the equation:

$$\sigma =\frac{\ln 2}{\pi RA}$$ (1)

Where σ is conductivity, R is resistivity and A is thickness of the films in centimeter (cm). The electrical conductivity increases with temperature. The electrical conductivity of the PVP/PAN/MWNTs nanocomposites can be analyzed according to the well-known Arrhenius equation:

$$\sigma ={\sigma }_0\exp \left[\frac{-E_a}{kT}\right]$$ (2)

where E_a is the conductivity activation energy, K is the Boltzman constant and σ_o is the pre-exponential factor including the charge carrier mobility and density of state. The semilogarithmic plots of ln σ vs. T⁻¹, illustrated in Fig. 5, are linear with conductivity activation energy E_a values of 92.2 meV and 43.6 meV for PVP/PAN and PVP/PAN/MWNTs, respectively. The corresponding values of activation energies are showed in Table 2.

Fig. 5.

Arrhenius plots of PVP/PAN/MWNT nanocomposites with 0 wt%., 0.71 wt% 1. 51 wt%, 1.90 wt% and 3.12 wt% of MWNTs.

Table 2.

Arrhenius and activation parameters for PVP/PAN/MWNTs.

wt% of MWNT	σo	Ea (meV)
0	0.0109	92.2
0.71	0.1541	79.5
1.51	0.1367	70.5
1.90	0.1557	69.4
3.12	0.0773	43.6

It was found that the electrical conductivity of the PVP/PAN (50:50 wt%) is 2.87 × 10⁻⁴ S/cm, which agrees well with that of Maitra et al [24]. With the variation of PAN, though the conductivity changes, the values lie in the range of 10⁻³ order of magnitude. However, changing the amount of PAN has considerable effect on mechanical properties. These observations indicate that PAN plays a great role in providing mechanical stability rather than conductivity [6, 34]. When the sample contains 3.12 wt% MWNTs, it get a highest electrical conductivity of 1.91 × 10⁻² S/cm.

Fig. 6 presents the relationship between conductivity of PVP/PAN/MWNTs nanocomposites and the volume fraction of MWNTs. The electrical conductivities of the PVP/PAN/MWNTs nanocomposites show typical concentration percolation behavior with the increase of MWNTs content as shown in Fig. 6. The electrical conductivity behavior above the percolation threshold can be adequately described by the scaling relation [35].

$$\sigma ={\sigma }_0{(\Phi /{\Phi }_0)}^t$$ (3)

where Ф is the volume fraction of the fillers and Ф_c is the percolation volume fraction. σ and σ_o are the conductivity of the composites and proportionality constant related to the intrinsic conductivity of the filler, respectively. The critical exponent (t) reflects system dimensionality of the composite. It follows a power-law dependence of approximately 1.6–2 in a three dimensional, and 1–1.3 in a two dimensional system [36, 37].

Fig. 6.

Electrical conductivity of PVP/PAN/MWNTs nanocomposites as a function of MWNTs volume fraction.

The conductivity values of PVP/PAN/MWNTs nanocomposites as a function of MWNTs volume fraction are presented in a log-log plot in Fig. 7. The best fitting to the experimental values resulted percolation volume fraction of 0.005 and the critical exponent of 1.62. The critical exponent of the nanocomposites system is in accordance with theoretical expectation and experimental observation, hinting formation of three dimensional percolating network at 0. 5 vol. % (0.71 wt%). Thus, the fitting result indicates a low percolation threshold at MWNTs loading of 0.71 wt%. Similar values of critical exponent were recently reported for epoxy/carbon nanotubes composites (t = 2.04) [38], for high density polyethylene/multi-walled carbon nanotubes composites (t = 1.67) [37] and for polyethylene/multi-walled carbon nanotubes composites (t = 1.77) [39]. The low value of Ф_c and high value of conductivity for PVP/PAN/MWNTs can be explained by high aspect ratio of the carbon nanotubes, good dispersion and distribution of the carbon nanotubes within the polymer composite.

Fig. 7.

Log-log plot of the conductivity as a function of Ф-Ф_c for the PVP/PAN/MWNTs nanocomposites. The solid lines are fits to a power law dependence of electrical conductivity on the MWNTs content (Eq. (3)).