Electrorheological behaviour of suspensions in silicone oil of doped polyaniline nanostructures containing carbon nanoparticles

Introduction

The use of external electric or magnetic fields to control the apparent viscosity of fluids has been attracting much attention in scientific and technological fields because of the potential use in active control of various devices in mechanics, biomedicine or robotics (Han et al., 2007; Sims et al., 2000; Tsuchiya et al., 2004). This type of fluids can vary their apparent viscosity reversibly responding to external electric or magnetic fields. They are often referred to as ‘smart fluids’ which include magnetorheological (MR) fluid, electrorheological (ER) fluid, liquid crystal (as a case of ER fluids) and ferrofluids (as a particular case of MR fluids). ER fluids consisting of polarizable particles dispersed in a non-conducting liquid are considered to be one of the most interesting and important smart fluids (Wen et al., 2008). It can be transformed reversibly in just milliseconds from a viscous state to an elastic state due to the disorder–order transition of particles under an applied external electric field, showing tuneable changes in the rheological characteristics. Under an applied electrical field, the polarizable particles orient along the field direction and connect themselves in chains or in columnar structures that span the space between the electrodes, which leads to tuneable changes in their rheological characteristics (e.g. yield stress, shear apparent viscosity, shear modulus, etc.). Their special rheological properties under the action of an external electric field makes the ER fluid possess potential uses to enhance the electric–mechanical conversion efficiency in mechanical devices such as clutches, valves, damping devices, ink jet printer, human muscle stimulator and mechanical sensor (Coulter et al., 1993; Hao, 2001).

For practical applications, a variety of polarizable particles, suspended in a non-conducting medium, have been investigated as potential ER materials, the particles being made of organic, inorganic, metallic, polymeric and hybrid materials (Choi and Jhon, 2009; Dong et al., 2014; McIntyre et al., 2012; Plachy et al., 2015a; Yin and Zhao, 2011). Due to its low density and abrasion to devices, organic ER systems have been widely investigated in the past decades. Polyelectrolytes and semi-conducting polymers are two kinds of important organic ER systems. The semi-conducting polymers including polyaniline (PANI) have been frequently considered as ER active materials, since the interfacial polarization, induced by the local drift of electrons or holes, provokes their ER effect. PANI is considered a promising conducting polymer with a range of advantages, such as outstanding electrical conductivity, good environmental stability, relatively low cost, facile synthesis and interesting electrical and optical properties (Riede et al., 2000; Stěnička et al., 2011; Yun et al., 2011). Just one form of PANI possesses high conductivity: emeraldine (oxidized form). This is a p-type organic semiconductor and its conductivity is favoured by dopant acid pHs (Ghosh et al., 2003)

Some studies about the ER effect of PANI hybridized with some inorganic materials, like vanadium oxide (Goswami et al., 2014), clays (Yoshimoto, 2005), nanotubes (Choi et al., 2005) and zeolites (Chattopadhyay et al., 2018), have been reported. For example, a recent study about the influence of silicone oil viscosity on the ER effect of PANI hybridized with graphene and tungsten nanoparticles has been published (Roman et al., 2018), proving the importance of this parameter. Furthermore, there are some studies about ER of PANI doped with HCl, which reveal the conducting properties of this polymer hybridized with multiwalled carbon nanotubes or colloidal graphene oxide (Park et al., 2005; Zhang et al., 2010). Nevertheless, the influence of different dopants or carbon nanoparticle concentration on the ER effect of PANI nanostructures has not been reported yet, which is the focus of this work and its novelty. There have been, however, some papers dealing with this topic using various concentrations of dopants for other conducting polymers or their oligomeric analogues (Stejskal et al., 2018). These works report that the pyrrole/p-benzoquinone systems show ER effect if appropriate concentrations of methanesulphonic acid are used. The same was also reported by Plachy et al. (2015b) for methanesulphonate salts of 2,5-(di-p-phenylenediamine)-1,4-benzoquinone.

The ER activity of the doped PANI including carbon nanoparticle–based systems has been thoroughly studied in this work. The plausible explanations behind the ER properties varying with different dopants, camphorsulphonic acid (CSA) and citric acid (CA), and with the carbon particle concentrations are presented in report. This system demonstrates its excellent characteristics as a promising candidate in future ER fluid systems.

Experimental section

Synthesis of materials and preparation of the samples

Reagents and materials

Aniline monomer (Sigma-Aldrich) was distilled under vacuum prior to use. Ammonium persulphate (APS; 99.99%, Sigma-Aldrich) and all other reagents like CSA (98%, Sigma-Aldrich), CA (≥99.5%, Sigma-Aldrich) and absolute ethanol (>99.99%, Alfa Aesar) were used as received. Carbon nanoparticles collected from candle shoot were used here.

Synthesis of materials

The polymerization was done in 1:1 (v/v) ethanol–water media. Ethanol helped to disperse the carbon nanoparticles within the precursor during the reaction and also slow down the polymerization process to some extent. The synthesis was carried out following a typical chemical oxidative polymerization of aniline at 0°C (ice bath) in the presence of APS as an oxidant and CSA or CA as a dopant. Aniline was added to 20 mL of ethanol–water solution and cooled at 0°C. A predetermined amount of carbon nanoparticles (1.1 and 2.2 mg for 1 and 2 wt% sample preparations) was dispersed in 10 mL of ethanol–water mixture by mechanical stirring and then APS (molar ratio of APS: aniline = 1:1) and CSA were added to it. The dopant CSA or CA was used in 1 wt% of samples. After that, this mixture was kept for cooling at 0°C for few minutes and then added dropwise to the precooled monomer solution with 30-s shaking. The polymerization was allowed to continue for 1 h without any kind of disturbance to the system. The blackish green precipitate of carbon nanoparticle–incorporated doped PANI was collected after filtration and repeatedly rinsed with deionized (DI) water and methanol. The products were finally dried in a vacuum oven at 60°C for 24 h to proceed on further investigations. The identification of the samples is presented in Table 1.

OPEN IN VIEWER

Table 1.

Nomenclature used for different samples as a function of PANI nanostructure concentration, carbon nanoparticles and the dopant.

Samples	%PANI nanostructures	%Carbon nanoparticles	CSA	CA
PANICA1-C1	1	1	No	Yes
PANICA2-C1	2	1	No	Yes
PANICSA1-C0	1	0	Yes	No
PANICSA1-C1	1	1	Yes	No
PANICSA1-C2	1	2	Yes	No

PANI: polyaniline; CSA: camphorsulphonic acid; CA: citric acid.

Preparation of the samples

All the samples were prepared by suspending the doped nanostructures in an insulating silicone oil followed by ultrasonication. The kinematic viscosities of the silicone oils were 50 and 20 cSt (Clearco Products Co., Inc.)

Results and discussion

Figure 1 presents the typical morphological features of the doped PANI sample containing carbon nanoparticles. It shows that the nanostructures are formed with particulate aggregates having rough surface.

OPEN IN VIEWER

Figure 1.

FESEM image of a typical doped PANI sample containing carbon nanoparticles.

The Raman spectra for PANICSA1-C1, PANICA1-C1 and carbon nanoparticles are shown in Figure 2. The Raman spectrum of carbon nanoparticles shows two broad characteristic bands at 1340 (D-band) and 1571 cm⁻¹ (G-band) which correspond to the presence of the amorphous disordered carbon structure of the nanoparticles. The Raman spectra of the carbon nanoparticle–percolated PANICSA and PANICA systems present some typical characteristic peaks of doped PANI with slightly shifted positions (Goswami et al., 2018; Wu et al., 2015). The peaks at 1166–1171 cm⁻¹ and 1589–1631 cm⁻¹ can be attributed to CH bending of the benzenoid ring and C=C stretching of the quinoid ring of PANI, respectively. The band situated at 1245 cm⁻¹ corresponds to CN stretching mode of the single bonds, and the one at 1484–1497 cm⁻¹ corresponds to CN stretching mode of the quinoid units indicating the presence of the doped PANI structure. The bands at 773–813 cm⁻¹ and at 1404–1407 cm⁻¹ are attributed to CH deformation and C=C stretching vibration of the quinoid ring, respectively. The peak at 1556 cm⁻¹ is already depicting the shift of the carbon particle–related band within the PANI matrix, showing the tight incorporation of the carbon nanoparticles through π–π synergy.

OPEN IN VIEWER

Figure 2.

Typical Raman spectra of carbon nanoparticles and the as-synthesized doped PANI with carbon particles inside.

Figure 3(a) and (b) shows the storage (G′) and loss (G″) modulus as a function of frequency for (a) PANICSA1-C1 and (b) PANICA1-C1 dispersed in 20 cSt (19 mPa s) silicone oil. These viscoelastic functions represent the elastic and viscous components of a viscoelastic material, respectively. G′ and G″ are markedly dependent of frequency without any electric field applied and G″ is higher than G′ in the entire frequency range studied for the PANICA1-C1 suspension. Hence, its behaviour is liquid like in the absence of any field. A similar nature is exhibited by the PANICSA1-C1 suspension. G′ and G″ increase with the electric field applied and both nanostructure suspensions change to a solid-like character upon the application of electric field. It is important to notice that for the same electric field applied G′ and G″ for the PANICSA1-C1 suspension are much higher than those for the PANICA1-C1 suspension. This may be due to the different acidity constants of the acids (k_CSA = 10^−1.2; k_CA = 10^−3.15) that provokes different conductivities and consequently polarization of the systems (Block et al., 1990). Since doped PANI is a p-type semiconductor, its conductivity is due to the high mobility of the charge carriers (holes). However, the only form of PANI that possesses high electrical conductivity is emeraldine salt. This form is favoured by low pH (Zuo et al., 1989); therefore, PANI doped with a stronger acid will be more conductive and more polarizable than that doped with a weaker acid. CA is weaker than CSA meaning that there will be less PANI in emeraldine form. The conductivities are 0.65 and 0.28 S m⁻¹ for PANICSA1-C1 and PANICSA1-C2, respectively, while the values are 0.03 S m⁻¹ for the PANICA1-C1 system and 1.5 × 10³ S m⁻¹ for carbon nanoparticles. In conclusion, even though the concentration of particles in the suspension may be the same, the polarization is higher in PANICSA1-C1, which leads to a higher ER effect in this case.

OPEN IN VIEWER

Figure 3.

Storage modulus (G′, solid points) and loss modulus (G″, open symbols) as a function of frequency under different electric fields applied, for suspensions of (a) PANICSA1-C1 and (b) PANICA1-C1.

Figure 4 shows the apparent viscosity curves for (a) PANICSA1-C1 and (b) PANICA1-C1 dispersed in 20 cSt (19 mPa s) silicone oil. All samples show shear thinning behaviour with a trend to reach a constant apparent viscosity value at high shear rates. An increase in apparent viscosity with the electric field applied is exhibited for the PANICSA1-C1 suspension for all the electric field strengths studied. However, the increase of apparent viscosity for the PANICA1-C1 suspension occurs only from 0.5 kV mm⁻¹ on. Since the electric field is perpendicular to the flow field, the particles reorient themselves in the direction of the electric field forming chains or column structures if the flow field is not strong enough to avoid it. This fact increases the apparent viscosity of the sample. At high flow field strengths, which means high shear rates, the flow field became dominant over the electric field leading to the destruction of the columns and to the corresponding apparent viscosity decrease and all the curves tend to the curve without any electric field applied. Furthermore, the shear rate for which a strong decrease in apparent viscosity is observed depends on the electric field strength, meaning that at higher electric fields higher shear rates are necessary to overcome the effect of the electric field.

OPEN IN VIEWER

Figure 4.

Flow curves under different electric fields applied, for the suspensions of (a) PANICSA1-C1 and (b) PANICA1-C1.

To better analyse the different flow behaviours of the samples, the apparent viscosity at 0.1 s⁻¹ is presented in Figure 5. There is a marked increase in apparent viscosity with electric field strength for 1 wt% suspensions doped with CSA (PANICSA1-C1) in comparison with those doped with CA (PANICA1-C1 and PANICA2-C1). The reason for this difference is the aforementioned different acidity constants of the acids. Even when doubling the carriers’ concentration in the case of PANICA suspension (PANICA2-C1), the ER effect is still much higher for the PANICSA1-C1 suspension, as shown in Figure 5.

OPEN IN VIEWER

Figure 5.

Influence of the dopant, particle concentration of PANICA suspensions and electric field strength on apparent viscosity at 0.1 s⁻¹ for suspensions of PANI nanostructures containing 1 wt% of carbon nanoparticles: PANICSA1-C1, PANICA1-C1 and PANICA2-C1.

Figure 6(a) shows the flow behaviour of PANICSA samples as a function of carbon nanoparticle concentration without the electric field applied (PANICSA1-C0, PANICSA1-C1 and PANICSA1-C2). The flow curves are quite similar, showing shear thinning behaviour with a trend to reach a Newtonian plateau for the highest shear rates. Hence, these samples possess the same flow properties without the electric field applied. However, the behaviour is different when an electric field is applied (Figure 6(b) and (c)). Figure 6(b) and (c) exhibits the flow properties of the PANICSA samples as a function of carbon nanoparticle concentration at 0.5 and 1 kV mm⁻¹, respectively. Although the form of the flow curves is the same, the values of apparent viscosity are really different at low and medium values of shear rate. At higher shear rates, the samples exhibited the same apparent viscosity. It is interesting to note that the addition of carbon nanoparticles provoked a decrease of the ER effect. This decrease was more marked for higher carbon nanoparticle concentrations. To better analyse the ER effect with carbon nanoparticle concentration, Figure 7 is presented. Figure 7(a) shows the apparent viscosity at 0.1 s⁻¹ for PANICSA nanostructures containing different concentrations of carbon nanoparticles (PANICSA1-C0, PANICSA1-C1 and PANICSA1-C2).

OPEN IN VIEWER

Figure 6.

Flow curves for the PANICSA1-C0, PANICSA1-C1 and PANICSA1-C2 suspensions (a) without electric field and (b) at 0.5 kV mm⁻¹ and (c) at 1 kV mm⁻¹.

OPEN IN VIEWER

Figure 7.

(a) Apparent viscosity at 0.1 s⁻¹ for suspensions of 1 wt% PANICSA nanostructures containing different concentrations of carbon nanoparticles (PANICSA1-C0, PANICSA1-C1 and PANICSA1-C2 suspensions) and (b) flow curves for 0.45 wt% carbon nanoparticle suspension as a function of electric field.

Figure 7(a) shows that there is a decrease in the ER effect with carbon nanoparticle concentration. This could be explained by the fact that carbon nanoparticles act as charge trapping centres within the conductive matrix of PANI. While travelling through the polymeric backbone by the hopping mechanism, the charge carriers are momentarily trapped within the carbon nanoparticles, thus restricting their normal path for conduction. It means that the polarizable nanoparticles are prevented from making stable columns, leading to a reduction in apparent viscosity (Gurram, 2017).

It could be argued that the synthesis and doping process of the PANI is not complete due to the presence of carbon nanoparticles; however, it is not so. Actually, the surface of the carbon nanoparticles becomes passivated by the –NH₂ group–rich aniline during in situ polymerization (Goswami et al., 2018) and the carbon–PANI forms a synergetic system through π–π transitions. According to the literature review, the conductivity of the conducting filler–loaded polymer depends largely on the content of fillers. Carbon particles exhibit the percolation phenomenon. However, there is a critical content depending on which the final C-PANI system shows its electrically conductive behaviour. Below this critical concentration, there is no continuous conductive network existing through the matrix due to the large separations among the fillers (Zhang et al., 2007). Thus, at such low carbon nanoparticle contents, the conductivity of the system depends mainly on the conductivity of the polymer matrix, which in this case is doped PANI, and therefore the conductivity depends on its doping level. The carbon nanoparticles behaving as islands within the matrix act as charge trapping sites for the hopping charge carriers. Accordingly, the polarizability of our composite ER suspension decreases in the presence of carbon nanoparticles which justifies the continuous decrease of ER activity with the increase of carbon nanoparticle concentration. This explanation is reinforced by Figure 7(b). Carbon nanoparticles show higher ER effect which is consistent with the high conductivity of pristine carbon. But, when they are incorporated in the PANICSA system (with a concentration below the critical one we mentioned before), the conductivity follows the charge trapping mechanism and hence the ER activity decreases.

The ER efficiency of PANICSA samples is shown as a function of electric field in Figure 8, for different carbon nanoparticle concentrations (PANICSA1-C0, PANICSA1-C1 and PANICSA1-C2). The efficiency of each fluid has been determined from the value of(τ − τ₀)/τ₀, where τ and τ₀ are the shear stresses with and without the electric field at a low shear rate (0.13 s⁻¹). The corresponding value of this parameter for each sample gives the relative increase in its shear stress (hence the apparent viscosity), with respect to the applied electric field. The efficiency relates to the utility of the material in the practical field (Goswami et al., 2014). There is an increase of efficiency with electric field strength for all the systems; however, a marked decrease in efficiency with carbon nanoparticle concentration is observed.

OPEN IN VIEWER

Figure 8.

Efficiency as a function of electric field for PANICSA1 samples containing 0, 1 or 2 wt% carbon nanoparticles.

Figure 9 shows the influence of silicone oil viscosity and electric field on the flow curves for PANICSA suspensions containing 2 wt% carbon nanoparticles (PANICSA1-C2). The flow curve of the 50 cSt system showed a higher limit Newtonian apparent viscosity than the flow curve of its counterpart. This fact is due to the fact that the apparent viscosity of the systems is related to the viscosity of the continuous phase at high shear rates. The same fact is observed under the application of 2 kV mm⁻¹ at high shear rates. By contrast, the apparent viscosity for the 50 cSt systems at low shear rates was slightly lower than the apparent viscosity for the 20 cSt system. To better distinguish the different trends, Figure 10 shows the apparent viscosity at 0.1 s⁻¹ for the PANICSA1-C2 sample as a function of continuous phase viscosity and the electric field applied. The increase of the continuous phase viscosity provoked a decrease in the ER effect. This fact is due to the increase of the difficulty in the formation of columns or chains when the oil viscosity is higher.

OPEN IN VIEWER

Figure 9.

Influence of silicone oil viscosity and electric field on flow curves for the PANICSA1-C2 suspension.

OPEN IN VIEWER

Figure 10.

Influence of silicone oil viscosity and electric field on apparent viscosity at 0.1 s⁻¹ for the PANICSA1-C2 suspension.

Conclusion

Novel ER fluids with doped PANI nanostructures containing carbon nanoparticles dispersed in silicone oil have been characterized using steady-state and low-angle oscillatory shear (SAOS) tests. The fluids showed increases in viscoelastic properties and apparent viscosity with the electric field applied, which are typical of ER fluids. The dopant (CSA/CA) of the PANI nanostructures is a key factor. The higher the acidity, the higher the ER effect. This is due to the concentration of emeraldine, the conductive form of PANI. In order to control the conduction mechanism within the PANI backbone, carbon nanoparticles were introduced as charge trapping centres. The addition of carbon nanoparticles provoked a decrease in the ER effect. This can be attributed to the fact that it is more difficult for these nanostructures to form columns in the presence of carbon nanoparticles because, acting as charge trapping centres, basically they defer the carrier conduction within PANI by the hopping mechanism. Consequently, the ER efficiency is higher for systems without carbon nanoparticles. Dynamic oscillatory tests in the linear viscoelastic regime support the results obtained in the steady shear measurements. The suspensions change their behaviour from fluid like to solid like, but in different degrees. The higher storage modulus value of the sample doped with CSA is due to the stiffer columnar structures that are formed in samples with higher emeraldine concentrations. Finally, the influence of nanostructure concentration and oil viscosity on the ER effect was studied. The increase of nanostructure concentration provoked an increase in the number of columns, leading to higher ER effect. By contrast, the increase in silicone oil viscosity reduces the ease of formation of columns provoking a reduction of the ER effect. Therefore, this work extends the knowledge about PANI-based nanostructures and demonstrates the important role of the dopant.