Electrical transport in titania nanoparticles embedded in conducting polymer matrix

1 Introduction

Polymer nanocomposites are a special class of materials and have a wide range of potential applications [1, 2] including sensors [3–6], membranes [7, 8], solar cells [9], biomedical and biomimic materials [10, 11], microwave absorption [12, 13], electrochromic devices [14], electrocatalysts for fuel cells [15] and structural materials [16–18] arising from their unique physicochemical properties, which are superior to conventional composites with filler size larger than 100 nm. The properties and performances of these polymer nanocomposites can be tuned to satisfy the desired applications through varying the filler material, size, shape and loading level in the polymer matrix.

The large surface area and the quantum size effect of the nanoparticles result in unusual physical properties of these materials. The properties of nanocomposites are quite different from the constituent materials due to molecular interactions. Moreover, the properties can be tailored by varying the composition. The main dificulty is the synthesis of inorganic nanosized fillers in the matrix of conducting polymers which are infusible and are not soluble in common solvents. The aggregation of nanoparticles with high surface energy also limits the preparation of nanostructured composites. Han and Armes [19] prepared a homogeneous nanocomposite of oxide materials using colloidal sol. Polymerization takes place in the presence of suspended colloid particles.

Conducting polypyrrole (PPY) is one of the most important and well-studied polymers owing to its ease of synthesis, reasonably good room temperature conductivity, stability in air and different potential applications. Among the various inorganic oxides, titanium dioxide (TiO₂) is a very fascinating semiconductor for wide applications in catalysis, dielectric ceramics, solar cells and optoelectronic devices [20, 21]. Composites of TiO₂ with different polymers such as conducting polyaniline [22–25], poly(phenylene vinylene) [26] and poly(methyl methacrylate) [27] have been studied in the last few years. Most of the studies concentrate on the optical properties of polymers surface modified TiO₂ nanoparticles. Growth of polymers in the presence of inorganic materials strongly affects the microstructure of the polymers. The morphology of polymers and the interface between polymers and nanoparticles determine the electrical properties of the nanocomposites. Impedance spectroscopy is an important technique to understand the charge transport mechanism and to find the nature of charge carriers in complex systems. Conductivity of TiO₂ is very low compared to PPY [28]. The change in conductivity with different compositions of PPY and understanding of the conduction process are important issues in nanocomposites. Transport and dielectric properties of the conducting polymer (PPY) containing nanosized semiconductor (TiO₂) have not yet been studied in detail. In this work, we present the transport and dielectric properties of PPY-TiO₂ nanocomposites at low temperature and high frequency.

2 Materials and methods

Pyrrole and ammonium peroxodisulfate (APS) are purchased from E. Merck (India). The monomer is vacuum distilled twice before use and kept in the dark prior to use and APS is used as received.

TiO₂ colloid is synthesized at room temperature by the acid hydrolysis of titanium isopropoxide (Sigma-Aldrich, India). In a typical synthetic procedure, 8 g of titanium isopropoxide is dissolved in 50 ml of 0.1(N) HCl (E. Merck, India) to get a clear solution. This solution is added dropwise to 500 ml of double distilled water under constant stirring. The pH of the resulting sol is adjusted using dilute NH₄OH (E. Merck, India) until it becomes 3. Thus the white colored colloidal sol formed is dialyzed until free from ions and stored in a 500 ml volumetric flask. This colloidal sol is used to prepare the nanocomposite samples.

For the preparation of the PPY-TiO₂ nanocomposites, 50 ml of TiO₂ colloid is taken at each time and the volume is reduced to about 20 ml on evaporation. A selected volume of pyrrole is then added to the colloidal sol under sonication to get TiO₂ nanoparticles impregnated with pyrrole having different weight ratios of colloid:monomer. After some time, APS solution maintaining a pyrrole:APS mole ratio of 1:1.25 is added dropwise under sonication. Immediate blackening of the solution indicates the formation of PPY. The solution is kept under sonication for about 1 h for complete polymerization followed by centrifugation at 10,000 rpm. The resulting nanocomposites come out as black solids which are washed thoroughly first with ethyl alcohol and then with distilled water several times. Finally, the composite samples are dried in a vacuum oven at 50°C for 24 h. Compositions of different nanocomposite samples with different TiO₂ (%) loadings studied are shown in Table 1.

3 Characterization

The particle content of the colloidal TiO₂ sol is measured by freeze drying followed by solvent evaporation at -35°C using a lyophilizer. The dried particles are weighed until a constant weight is achieved.

The particle size of the bare colloid and the nanocomposites and the nature of interaction between the conducting and insulating components are determined from scanning electron microscopic (SEM; Model: JEOL-JSM 6700F) and transmission electron microscopic (TEM; Model: JEOL, HRTEM JEM 2010) studies. Infrared spectra of the bare polymer and the nanocomposites samples pelletized with KBr are performed using a Fourier transform infrared (FTIR) spectrometer (Perkin-Elmer Model 1600). X-ray diffraction (XRD) patterns of the nanocomposites are performed using a Philips diffractometer (PW 1710) using Cu Kα radiation.

The temperature-dependent direct current (dc) conductivity is measured by the standard four-probe method (Keithley Instrument, Model 2000). The complex dielectric constants are obtained from the measurements of capacitance (C) and dissipation factor (D) by a 4192A Agilent impedance analyzer (Hewlett Packard) up to the frequency of 1.6 MHz at different temperatures. The real part of the dielectric constant ε₁ is evaluated by the relation C=ε₁ A/t, where A is the area and t is the thickness of the sample. The imaginary component is calculated from the dissipation factor, ε₂=D ε₁. The electrical contacts are made by silver paint (Supplied by Acheson Colloiden B.V., Netherland).

4 Results and discussion

Figure 1 shows the characteristic peaks of XRD of the nanocomposite sample (S1) with highest content of TiO₂. The main peaks at 2θ=25.3° (101), 37.9° (004), 47.7° (200), 53.3° (105) and 62.9° (204) which are characteristics of TiO₂ are also present in the composite. The absence of peak at 2θ=27.4° and 36.1° rules out the possibility of the existence of a rutile phase of titania. The crystallite size of the TiO₂ nanoparticles in the composites is calculated following the Scherrer equation [29]:

Figure 1:

X-ray diffraction pattern of TiO₂-polypyrrole (PPY) nanocomposite sample S1.

where K=0.89, D represents crystallite size (nm), λ is the wavelength of Cu Kα radiation and β is the corrected value at full width of half maxima of the diffraction peak. To calculate the average diameter, 2θ=25.3° (101 face) is chosen, which is the characteristic peak of TiO₂; it comes out to be 10 nm, which is consistent with that obtained from TEM studies.

The transmission (TEM) and scanning (SEM) electron micrograph of the nanocomposite sample S1 is shown in Figure 2A, B and C, respectively. Figure 2A shows the lower magnification image of the sample S1, which indicates that the nanoparticles are well dispersed in the polymer matrix and are of spherical shape with uniform diameter lying in the range from 20 nm to 30 nm. The high resolution TEM (HRTEM) of the same sample is shown in Figure 2B. This shows the lattice image from a TiO₂ nanoparticle in the background of a PPY matrix. The lattice spacing is found to be 0.148 nm, which corresponds to (204) plane in TiO₂. After the formation of the composites, the particles (dark shaded) are found to be encapsulated into PPY (light shaded) chains. Therefore, it can easily be concluded that the colloid particles are not simply mixed up or blended with the polymer, they are rather entrapped inside the PPY chains. The HRTEM image shows the size of the TiO₂ nanoparticle in the sample S1 to be 20 nm, which indicates polymers are formed on the surface of individual nanoparticles. SEM of the nanocomposite sample as presented in Figure 2C reveals that the grain sizes are of the order of 100–200 nm. The size decreases with the increase of TiO₂ content in the composites. Moreover, the grain becomes more uniform with the increase of TiO₂ content.

Figure 2:

Transmission electron microscopy (TEM) micrograph of the sample S1. (A) Lower magnification TEM image, (B) high resolution lattice image and (C) scanning electron microscopy (SEM) micrograph of the sample S1.

Figure 3 shows FTIR spectra of PPY and the nanocomposite samples S3 and S4, respectively. The peak at 1541 cm^-1 corresponds to typical pyrrole rings vibration. The peaks at 1300 cm^-1 and 1170 cm^-1 are attributed to =CH in plane vibration and peaks at 784 cm^-1 and 891 cm^-1 due to =CH out of plane vibration. The characteristic band of PPY is shifted to a higher wavenumber in the nanocomposites. This indicates that there is a strong interaction between PPY and TiO₂ nanoparticles.

Figure 3:

Fourier transform infrared (FTIR) spectra of (A) pure polypyrrole (PPY), (B) nanocomposite S2 and (C) nanocomposite S3, respectively.

Room temperature conductivities σ (RT) for various compositions are shown in Table 1. The dc conductivity increases with the increasing content of PPY. The enhancement of σ(RT) is about two orders of magnitude for the highest fraction of PPY. The values of temperature dependence of conductivity for four samples are shown in Figure 4. Temperature variations of conductivity clearly show that all samples are semiconducting in nature. The temperature dependence of conductivity σ(T) of disordered semiconducting materials is generally described by the Mott’s variable range hopping (VRH) model [30]:

Figure 4:

Temperature variation of direct current (dc) conductivity of the nanocomposites S1, S2, S3 and S4, respectively. The solid lines are fits to Eq. (2).

where σ₀ is the high temperature limit of conductivity and T₀ is Mott’s characteristic temperature associated with the degree of localization of the electronic wave function. The exponent γ=1/(1+d) determines the dimensionality of the conducting medium. The possible values of γ are 1/4, 1/3 and 1/2 for three (3D), two (2D) and one (1D) dimensional systems, respectively. The exponent q in pre-exponential factor is 1/2 in Mott’s three dimensional VRH. The plot of ln(σ[T]T^1/2) against T^-1/4 indicates that 3D charge transport occurs in the PPY-TiO₂ nanocomposites. The values of Mott’s characteristic temperature T₀ and the pre-exponential factor σ₀ are obtained from the slopes and intercepts of Figure 4 and are given in Table 1. The values of T₀ are very sensitive and decrease with increasing PPY content. A larger T₀ implies stronger localization of the charge carriers, with the increase of resistance. In the case of nanocomposites, T₀ increases with increasing particle loading indicating enhanced localization of the charge carriers. These are in good agreement with the theoretical observation.

The ac conductivity of samples S2, S3 and S4 are investigated. The variations of ac conductivity as a function of frequency at different temperatures are shown in Figure 5 for the sample S4. It is seen that σ(ω) remains constant at a low frequency. The conductivity starts to increase from dc value after a certain characteristic frequency ω₀ known as onset frequency. The extra contribution to conductivity comes from capacitive regions which provide less impedance at higher frequency. Figure 5 reveals that ω₀ is strongly temperature-dependent.

Figure 5:

Frequency dependence of conductivity at different temperatures of the sample S4.

A general feature of amorphous semiconductors and disordered systems is that the frequency-dependent conductivity σ(ω) obeys a power law with frequency. The total conductivity σ(ω) at a particular temperature over a wide range of frequencies can be expressed as:

where σ_dc is the dc conductivity and A is a constant depending on temperature. The frequency exponent lies between 0 and 1 and is given by S=d lnσ(ω)/d lnω. The values of s at each temperature has been calculated from the slope of log(σ_ω-σ_dc) vs. log ω plot. The estimated values of S are between 0.4 and 0.9 for all samples, as shown in Figure 6.

Figure 6:

Frequency exponent(s) vs. temperature (T) for the samples S2, S3 and S4, respectively. The solid lines are fits to Eq. (4).

Temperature dependences of S for three samples are shown in Figure 6. The value of frequency exponent S increases with increase of temperature. This behavior is only observed in small polaron tunneling models for conduction process [31]. A charge carrier along with local lattice distortion forms a polaron. In the case of a small polaron, distortion clouds do not overlap due to the localized behavior of the polaron. The temperature dependence of S based on this model is:

where W_H is the activation energy for polaron transfer, τ₀ is the characteristic relaxation time for small polaron tunneling and k is Boltzman constant. The values of W_H and τ₀ are obtained by the best fitted parameters to Eq. (4) at a frequency of 100 KHz and are given in Table 2. Figure 6 exhibits that experimental values of S are in good agreement with the theoretical model.

The tunneling distance [31] at a particular frequency and temperature is:

The ac conductivity [31] for small polaron tunneling is:

where L is the localization length C=π⁴/24, e is the electronic charge and N(E_F) is the density of states at Fermi level. The best fitted values of W_H and τ₀ are used to calculate R_ω and N(E_F) from Eqs. (5) and (6), assuming L=1.2 Å and are shown in Table 2. Tunneling distance increases and density of states decrease with increase of PPY content.

The relative dielectric constants (ε₁) as a function of frequency at room temperature for different compositions are presented in Figure 7. The magnitude and the frequency dependence of ε₁ are strongly influenced by the content of TiO₂ nanoparticles. An interesting fact is that a very high dielectric constant of about 13,000 at room temperature is found for the sample (S1) with highest content of TiO₂ (PPY~70.7%, TiO₂~29.3% as in Table 1). In the case of more conducting samples S4, the value of ε₁ reduces to about 6000, but remains almost unchanged within the measured frequency region. The values of ε₁ for the different samples at a fixed frequency (30 KHz) are given in Table 1.

Figure 7:

Real part of the dielectric constant vs. frequency at room temperatures for different samples.

The temperature variation of ε₁ for S1 is shown in Figure 8. A strong temperature dependence of dielectric constant at low frequency is observed at higher temperatures. A step-like increase in ε₁ is observed in the lower frequencies region. Figure 8 also indicates a weaker temperature dependency of ε₁ at lower temperatures. The values of ε₁ decrease sharply above some frequency depending on temperature. The resonating behavior in ε₁ is found at a few MHz region around room temperature. The maximum value of dielectric constant in TiO₂ is about 80. The value of ε₁ for PPY is approximately 570. The present observation of ε₁ is remarkable, as it is about 20 times larger than constituent materials.

Figure 8:

Frequency dependence of the real part of relative dielectric constant (ε₁) at different temperatures for the sample S1.

The dielectric loss spectra ε₂ as a function of frequency are shown in Figure 9 for different temperatures. The spectra exhibit broad peaks at low temperatures. The peak frequency shifts to a higher frequency with the increase of temperature. The complete loss spectra at higher temperatures are not observed due to our experimental limitation. The relatively large width and asymmetrical nature of peaks at a low temperature suggest the non-Debye behavior of the dielectric relaxation process. The dielectric behavior has been described by Havriliak-Negami (HN) function [32]:

Figure 9:

Frequency dependence of imaginary part of dielectric constant at different temperatures of the sample S4. The solid lines are fits to Eq. (7).

where τ is the average relaxation time which is given at the frequency of maximum dielectric loss. The difference (ε_s-ε_∞) is known as the dielectric relaxation strength. The parameter α describes the distribution of the relaxation time of the system. Debye relaxation is obtained for α=1 and β=1. The deviation of α from 1 indicates the broad distribution of relaxation time in the spectrum. Figure 9 shows that the experimental data are reasonably good fitted with the calculated value of HN function. The best fitted parameters are shown in Table 3. The values of α and β are different from unity which imply non-Debye relaxation process at low temperature. Temperature induces the Debye process as confirmed from the values of α and β at T≥178 K.

The relaxation time, τ at different temperatures is determined from the reciprocal of the peak frequency. The Arrhenius plot of lnτ against 1/T for the sample S4 is shown in Figure 10. A straight line behavior is obtained. Thus, temperature dependence of relaxation time of loss is given as τ∝exp(-E/kT), E is the activation energy of the dielectric process and k is the Boltzmann constant. The slope of the best fitted straight line gives the activation energy of 63 meV.

Figure 10:

Arrhenius plot of dielectric relaxation time vs. frequency at different temperatures for the sample S4.

High dielectric constant (k) materials have received increasing interest in recent years as they are attractive as potential materials for various applications including gate dielectrics [33], high charge-storage capacitor [34] and electroactive materials [35]. For instance, high dielectric constant and low dielectric loss materials are imperative to realize the real applications of an embedded capacitor, which is one of the emerging and important technologies for electronic packaging to provide the advantage of size reduction and system performance enhancement. A wide variety of materials have been extensively investigated as candidates for this application, such as perovskite oxides [36], ferroelectric ceramic [37], ferroelectric ceramic/polymer composite [38] and conductive filler/polymer composite [39]. Conductive filler/polymer composite materials, identified as a conductor-insulator percolative system, have been recognized as a promising method to achieve high k.

Constant efforts are made to increase this value for more technological applications of polymers. The dielectric constant of polymeric systems can generally be enhanced by adding high dielectric constant particulates. Recently it has been observed that the large dielectric constant can be obtained near the percolation threshold of the composite [40].

The nanocomposite consisting of grain and a more insulating grain barrier make the system heterogeneous. The conductivity of TiO₂ nanoparticles within the grain is less than that of the conducting PPY. As a result, the accumulation of charges is greater with the increase of TiO₂ concentration. The interfacial polarization arising from different conductivities of grain and grain boundary may be a source for the large dielectric constant. The large dielectric constant caused by macroscopic heterogeneity originates from Maxwell-Wagner type interfacial polarization [41, 42]. The dielectric constant of the nanocomposite increases with increase of TiO₂ content. The encapsulated TiO₂ nanoparticles play a significant role in enhancing the dielectric constant.

The complex dielectric constant can be calculated from the relation, ε^*=1/(iωC₀Z^*). The static dielectric constant of such a system consisting of different conductivities can be expressed as [43]:

where C₀=ε₀S/t is the geometrical capacitance. For the present materials, R_gb>>R_g so the static dielectric constant can be approximated from Eq. (5):

The interfacial polarization arising from different conductivities of grain and grain boundary may be the source for the large dielectric constant.

5 Conclusion

Transport properties of PPY-TiO₂ nanocomposites are studied at low temperature and high frequency up to 1.6 MHz. The variation of dc conductivity with temperature implies 3D VRH conduction processes. Ac conductivity spectra exhibit two distinct regions – constant conductivity at low frequency and conductivity dispersion at high frequency. A small polaron tunneling mechanism is used to interpret ac conduction. Both the interface and the nanosize of TiO₂ may play important roles in magnifying the dielectric properties. The exact physical origins of the unusual dielectric constant remain still unclear. However, the large dielectric constant of the nanocomposite provides promising materials for the applications in the field of micro actuators, dynamic random access memory and metal oxide semiconductor devices.