Introduction

Lamellar nanocomposites consisting of layered inorganic compounds and conducting polymers display novel properties which result from the molecular level interaction of two dissimilar chemical components [1–6]. Intercalative polymerization of aniline in α-RuCl3 host has been reported recently [1, 2]. The insertion of aniline into α-RuCl3 has been executed from a solution of aniline in acetonitrile. It has been proven that polyaniline was formed between the RuCl3 layers, which are composed of hexagonal sheets of Ru atoms sandwiched between two hexagonal sheets of Cl atoms with ABC stacking. In the Ru sheets, one-third of the positions are left unoccupied [1, 2, 7–13]. The RuCl3 is a strongly oxidizing host which can take up the electrons from the aniline leading to the formation of polyaniline (PANI). Simultaneously, a fraction of Ru3+ atoms is reduced to Ru2+, resulting in a mixed valence compound [1, 2]. The host material will have a negative charge, and RuCl −3 sites can act as counterions for anilinium cations and charged PANI in the nanocomposite, (PANI) z+ x (RuCl3) z− y . The X-ray diffraction pattern of the samples revealed that the structure of the inorganic host was preserved; however, a Δd=0.62 nm increase in the separation of the RuCl3 layers occurred [1, 2]. It has to be mentioned that the interlayer spacing also increases as a consequence of the intercalation of the hydrated ions during the reduction of α-RuCl3, e.g. Δd=0.54 nm (Li+) or 0.566 nm (Na+) [10, 11]. The new (PANI) x (RuCl3) y nanocomposite has also been characterized by thermogravimetric analysis, infrared spectroscopy, magnetic measurements as well as electrical and ionic conductivities and thermopower measurements [1, 2]. It has been established that the charge transport—which occurs by electron hopping between the ruthenium ions in the mixed-valent compound—is substantially enhanced by the presence of the conductive polymer. The results of the thermopower study indicate a bulk metal-like conductivity which is controlled by the conductive polymer. (PANI) x (RuCl3) y shows a room temperature conductivity of ca. 1 S cm−1 [2]. It was suggested that the combination of the high conductivity of polyaniline [14–17] and the wide-ranging catalytic properties of RuCl3 [7–9, 18–20] could provide new materials with valuable electrocatalytic properties [1, 2].

A logical continuation of these efforts is the study of the redox behaviour of this nanocomposite by electrochemical methods including the investigation of the ionic charge transport processes by nanogravimetry since the oxidation state and consequently the properties of the nanocomposite can be tuned electrochemically. In this article, we report the results on the behaviour of (PANI) x (RuCl3) y nanocomposite prepared in different ways. The microcrystals were immobilized at gold surfaces [21–23] and studied by electrochemical quartz crystal microbalance technique.

Experimental

Black-coloured RuCl3 microcrystals, HCl, NaCl, aniline (Merck, analytical grade) and doubly distilled water were used. In the EQCM measurements, 10 MHz AT-cut crystals coated with gold were used. The detailed description of the apparatus and its calibration have been published in our previous papers [24–27]. The geometrical and piezoelectrically active area of the working electrode was 0.4 cm2. A Pt wire was used as a counter-electrode. The reference electrode was a saturated sodium calomel electrode (SCE). All potentials are referred to SCE. The (PANI) x (RuCl3) y samples were prepared in three different ways. First, the procedure described in [1, 2] was used, i.e. 0.1 g RuCl3 was added to 10 cm3 of 4% aniline/acetonitrile solution; the suspension was stored for 5–7 days and occasionally stirred. The product [labelled (PANI) x (RuCl3) y (A)] was washed with acetonitrile and dried. Second, the same procedure was followed except pure aniline was added to the RuCl3 microcrystals and HCl was used to wash the resulting material (labelled B). Third, nanocomposites A and B attached to gold were investigated by electrochemical quartz crystal microbalance in the presence of HCl or NaCl solutions and then aniline was added to the solution. The incorporation of aniline was detected by measuring the frequency change. The nanocomposite formed (marked as C) was investigated further on when practically no more frequency decrease could be observed.

The microcrystals were attached to the gold surface by placing the material on the electrode [24–27] and then, making use of a drop of water, the microcrystals were fixed. Although RuCl3 and (PANI) x (RuCl3) y microcrystals were found to be insoluble in water, using this procedure the crystals can be “glued” to the metal. This method of preparing the electrode produces a randomly distributed ensemble of microcrystals on gold; however, the distribution of microcrystals was more or less uniform. Although the requirements (uniform and homogeneous surface layer) for the application of Sauerbrey equation are not perfectly met, on the basis of measured frequency values (Δf) a rough estimation can be done. The relative values of Δf obtained for the incorporation of aniline or different ions and solvent molecules, however, should be approximately correct. Therefore, Sauerbrey equation was used for the calculation of the surface mass changes (Δm) from the frequency changes (Δf), with an integral sensitivity [C f=(1.9±0.2)×108 Hz cm2 g−1] that was determined in separate experiments [27]. The amount of RuCl3 and (PANI) x (RuCl3) y immobilized on the surface (0.5–15 μg) was estimated by measuring the crystal frequency before and after the deposition in dry state. Solutions of different concentrations of HCl and NaCl were used. The measurements were carried out at 25°C. All solutions were purged with oxygen-free argon and an inert gas blanket was maintained throughout the experiments.

An Elektroflex 453 potentiostat and an Universal Frequency Counter TR-5288 connected with an IBM personal computer were used for the control of the measurements and for the acquisition of data. (PANI) x (RuCl3) y nanocomposites as pressed KBr pellets were investigated by infrared (FTIR) transmission spectroscopy using a Bruker IFS-55 Fourier Transform Spectrometer.

Results and discussion

Figure 1 shows the cyclic voltammograms obtained for RuCl3 and (PANI) x (RuCl3) y (B) samples attached to a gold electrode and studied in the presence of 0.5 mol dm−3 HCl.

Fig. 1
figure 1

Cyclic voltammograms obtained at different scan rates a v=50 mV s−1 and b v=5 mV s−1 for Au|RuCl3 (curve 1) and Au|(PANI) x (RuCl3) y (B) (curves 2, 3, 4). Electrolyte: 0.5 mol dm−3 HCl

Full size image

In these experiments, first pure α-RuCl3, then (PANI) x (RuCl3) y (B) prepared by 1-week-long soak of α-RuCl3 microcrystals in aniline were immobilized at the gold surface. The nanocomposite was washed with 0.5 mol dm−3 HCl before use. A comparison of the cyclic voltammograms displayed in Fig. 1a reveals that the oxidation of Ru2+ to Ru3+ becomes easier since wave II moves into the direction of lower potentials while the reduction process remains unaltered. It may be related to the presence of polyaniline, which is conducting in this potential region, and probably enhances the charge transfer processes. The waves belonging to the leucoemeraldine \( (\hbox{LE}) \rightleftarrows \hbox{emeraldine} \, (\hbox{E})\) transition are clearly seen in Fig. 1a (waves III and IV). We can assign these waves to the redox transformations of polyaniline since the peak potentials are close to those observed for PANI films in this media [14, 28]. Figure 1b shows the cyclic voltammograms obtained at a slow scan rate in the whole potential region and where the redox transformations of RuCl3 play no role, i.e. the response of PANI can be seen separately. An even more direct proof regarding the formation of polyaniline is furnished by the IR spectra. The IR spectrum obtained for the sample used in the experiments described above is shown in Fig. 2. Almost all peaks in the IR spectrum of (PANI) x (RuCl3) y presented in Fig. 2 are associated with the emeraldine salt form of polyaniline. In the previous publications similar spectra have been reported [2, 29–36]. The characteristic vibration bands can be assigned as follows: benzoid ring (1,561 cm−1), quinoid ring (1,488 cm−1), –C–N stretch (1,305 and 1,241 cm−1), –C–H in plane (1,146 cm−1), –C–H out of plane (815 cm−1).

Fig. 2
figure 2

Infrared spectra of (PANI) x (RuCl3) y (B)

Full size image

The electrode used in the experiments presented in Fig. 1 was soaked in a solution containing 0.4 mol dm−3 aniline and 0.2 mol dm−3 HCl for 3 h (nanocomposite C). The cyclic voltammogram shown in Fig. 3a (curve 1) clearly indicates the additional deposition of PANI which means that RuCl3 was still able to oxidize aniline and enhance the formation of PANI. In order to ascertain that the redox behaviour originated from PANI, some layers of PANI were electropolymerized on top of the crystals as shown in Fig. 3b. The cyclic voltammetric curves obtained before (curve 1) and after (curve 2) the electropolymerization are displayed in Fig. 3a. The comparison reveals that there is no doubt that the original response is also due to the PANI redox transformations.

Fig. 3
figure 3

Cyclic voltammograms of Au|(PANI) x (RuCl3) y (C) electrodes a after soaking the electrode used in Fig. 1 in a solution containing 0.4 mol dm−3 aniline and 0.2 mol dm−3 HCl for 3 h (curve 1); the cyclic voltammogram taken after the electropolymerization shown in (b) (curve 2). Electrolyte: 0.5 mol dm−3 HCl, scan rate: 5 mV s−1. b Electropolymerization of aniline by potential cycling between −0.2 and +0.8 V versus SCE on the (PANI) x (RuCl3) y (C) electrode used in the experiment shown in curve 1 of a. Solution: 0.4 mol dm−3 aniline + 0.2 mol dm−3 HCl; scan rate: 50 mV s−1

Full size image

The electrochemical activity of PANI decreases with increasing pH [37, 38], and at pH>5—except for self-doped films [35, 36, 39–43]—no redox response can be observed. Therefore, we have tried whether the (PANI) x (RuCl3) y composite behaves as a self-doped system or not. Figure 4 shows the voltammograms of the nanocomposite B in the presence of 0.5 M HCl and 0.5 M NaCl, respectively. Although both waves belonging to Ru3+→Ru2+ and LE → E transition, respectively, move into the direction of higher potentials, it is clearly seen that the electrochemical activity (see waves III and IV) of PANI was preserved. The sharp pair of waves at low potentials is a typical response of α-RuCl3 in neutral salt solutions [44]. Its nature has been discussed recently [44]; however, the presence of PANI certainly enhances this process.

Fig. 4
figure 4

Cyclic voltammetric curves obtained for a Au|(PANI) x (RuCl3) y (B) electrode in the presence of (1) 0.5 mol dm−3 HCl and (2) 0.5 mol dm−3 NaCl. Scan rate: 5 mV s−1

Full size image

The scan rate dependence shown in Fig. 5 attests that the redox transformations of both ruthenium and polyaniline are fast; however, the mass changes accompanying the redox reactions somewhat decrease with increasing scan rate. All the waves behave as surface waves at low scan rates since the peak currents are proportional to the scan rate; however, an interplay between the surface and diffusional behaviour can be observed at v>ca. 50 mV s−1, and eventually a diffusional response develops.

Fig. 5
figure 5

The effect of scan rate on a the cyclic voltammograms and b the simultaneously detected EQCM frequency curves obtained for a Au|(PANI) x (RuCl3) y (B) electrode. Scan rates are (1) 5, (2) 20, (3) 50 and (4) 100 mV s−1. Electrolyte: 0.5 mol dm−3 HCl

Full size image

When aniline was added to the solution at open circuit (E=0.21 V), a fast frequency decrease was observed as shown in Fig. 6. It is somewhat surprising since the sample attached to the gold surface had been soaked in an aniline-containing solution and showed the presence of PANI (see Figs. 1, 4, 5). The amount of aniline adsorbed/absorbed was 2.6×10−7 g, which means that the resulting layer contains ca. 1 mol aniline/1 mol RuCl3. The mass increases/decreases in the course of subsequent potential steps were in accordance with those observed in cyclic voltammetric experiments. During the potential step investigations the potential region where aniline electropolymerization could have occurred was avoided. In Fig. 7 the cyclic voltammetric and microgravimetric responses before (nanocomposite B) and after the addition of aniline (nanocomposite C), respectively, are compared. The anodic part of the voltammogram became more drawn-out, and the cathodic waves shift to higher potentials. Even more drastic changes of the EQCM frequency curves can be observed. The steep frequency increase at the reduction wave of Ru3+ disappeared and a rather reversible frequency variation can be observed. Since this frequency increase has been assigned to the desorption of water molecules [24, 44], we may conclude that aniline molecules replaced water molecules within the interlayer space. It should be mentioned that potential cycling in the presence of aniline and HCl for a longer period of time—in the potential region where electropolymerization does not take place—leads to an excessive insertion of aniline (anilinium chloride), which eventually decreases the electrochemical activity of the layer. It may be related to the diminishing charge transfer between the RuCl3 layers.

Fig. 6
figure 6

Sorption of aniline in the (PANI) x (RuCl3) y (B) layer attached to the gold surface. The moment of the addition of aniline to the 0.5 mol dm−3 HCl supporting electrolyte is indicated by an arrow. Frequency changes (dotted line) and current responses (continuous line) in the course of the open circuit and the subsequent potential step experiments are shown

Full size image
Fig. 7
figure 7

A comparison of a the cyclic voltammetric and b the simultaneously obtained EQCM frequency responses before (nanocomposite B, curve 1) and after the addition of aniline (nanocomposite C, curve 2) the experiment of which is presented in Fig. 6. Electrolyte: 0.5 mol dm−3 HCl, scan rate: 5 mV s−1

Full size image

The behaviour of the (PANI) x (RuCl3) y samples produced by different treatments has also been compared. The voltammetric and EQCM responses of the layers prepared in aniline (nanocomposite B) (presented in Figs. 4, 5, 7) and in aniline/acetonitrile solution (nanocomposite A), respectively, are shown in Fig. 8. The latter (PANI) x (RuCl3) y (A) nanocomposite displays a more reversible behaviour especially regarding the reoxidation of Ru2+ (wave II). The mass change is also more regular in as much as there are no crossed curves. Furthermore, at lower potentials a mass increase and decrease during oxidation and reduction, respectively, can be observed, while it was just the opposite in the case of the nanocomposite layer prepared without acetonitrile. Another important feature is the much higher frequency change at the \( \hbox{Ru}^{3+} \rightleftarrows \hbox{Ru}^{2+}\) transitions. In order to clarify the nature of the mobile species causing the mass increase and decrease phenomena the Au|(PANI) x (RuCl3) y (A) electrode has been studied in perchloric acid media also. As seen in Fig. 9 there is only a slight change in the voltammetric response and—as expected—a higher mass change occurs in the presence of ClO −4 ions in comparison with that of Cl− ions. However, the variation is much below the theoretical M(ClO −4 )/M(Cl−)=2.8 molar ratio. From the slope of the Δf(Δm) versus Q plots the apparent molar masses were calculated, which gave M app(ClO −4 )=124±15 and M app(Cl−)=94±15, respectively.

Fig. 8
figure 8

A comparison of a the cyclic voltammetric and b the EQCM frequency responses of two Au|(PANI) x (RuCl3) y electrodes. The nanocomposites were prepared in RuCl3–aniline (nanocomposite B, curve 1) and in RuCl3–aniline–acetonitrile (nanocomposite A, curve 2) suspensions, respectively. Electrolyte: 0.5 mol dm−3 HCl, scan rate: 5 mV s−1

Full size image
Fig. 9
figure 9

The effect of the anions on a the cyclic voltammetric and b EQCM responses. Au|(PANI) x (RuCl3) y (A) prepared in aniline/acetonitrile solution in contact with (1) 0.5 mol dm−3 HClO4 (continuous line) and (2) 0.5 mol dm−3 HCl (dotted line), respectively. Scan rate: 5 mV s−1. c The frequency is plotted as a function of the charge consumed (Q)

Full size image

Although this discrepancy may be explained by the higher hydration of chloride ions, i.e. chloride ions transfer more water molecules, it is likely that the phenomena observed are not related to the sorption and desorption of anions. The results obtained for α-RuCl3 microcrystals indicate that at peak I a phase transition occurs which is accompanied by the desorption of water molecules, and cations enter the solid phase during the reduction [22, 44]. The different behaviours of nanocomposites A and B (Fig. 8) as well as the effect of excess aniline support the idea that the structural changes of α-RuCl3 play a decisive role in the behaviour of the layer including the sorption/desorption of mobile species. The most versatile behaviour is expected concerning the incorporation of neutral molecules since it is only indirectly affected by the charging state of the solid phase. Nevertheless, we should not exclude anion sorption at high potentials since the RuIII Cl3 cannot act as counterions. At low potentials the reduced PANI cannot compensate the negative charge of the reduced RuII Cl −3 sites; therefore, the incorporation of cations in partially hydrated form is expected. In acidic solutions PANI may posses positive charge due to its protonation (protonated LE form) [28, 45, 46]; however, it is sufficient to maintain the electroneutrality only in the case when x = y and PANI is fully protonated. From the charge consumed in the course of the cyclic voltammetric experiments a rough estimation can be made for the x/y ratio in the (PANI) x (RuCl3) y nanocomposites. For the nanocomposites A and B this ratio is 0.5–0.9 and 0.3–0.8, respectively, depending on the preparation conditions. By the treatment in aniline–HCl solutions, x/y∼1 can be achieved (see Fig. 3a); however, further incorporation of anilinium ions eventually destroys the structure of the microcrystals that leads to the decrease of the electroactivity. In HCl electrolytes only H+ (or H3O+) ions can enter the nanocomposite or alternatively H+ ions insertion occurs and water molecules leave the film. In the first case, only a rather small frequency decrease is expected, while in the second case even frequency increase can be observed depending on the extent of counterflux of the mobile species. On the other hand, in NaCl solutions a more distinct mass increase has to be observed due to a factor of 23 in the molar masses of Na+ as compared to H+. As seen in Fig. 10 it happens indeed, although the picture is somewhat complicated due to a second reduction process of RuCl3 in neutral media.

Fig. 10
figure 10

Cyclic voltammetric (a) and EQCM responses (b) obtained for (PANI) x (RuCl3) y (A) nanocomposite in the presence of 0.5 mol dm−3 HCl (1) and 0.5 mol dm−3 NaCl (2), respectively. Scan rate: 5 mV s−1

Full size image

Conclusions

The results of the electrochemical microgravimetry support the previous findings [1, 2] that aniline molecules can be intercalated in the lamellar structure of α-RuCl3. Ru3+ ions oxidise aniline molecules resulting in the formation of polyaniline monolayers within the gallery space of RuCl3. The properties of the (PANI) z+ x (RuCl3) z− y nanocomposite somewhat depend on the method of preparation; however, in all cases the electrochemical responses clearly show the redox responses of both leucoemeraldine \(\rightleftarrows\) emeraldine and \(\hbox{Ru}^{3+} \rightleftarrows \hbox{Ru}^{2+}\) transitions. The results of electrochemical nanogravimetric studies can be explained by the sorption/desorption of ionic species and water molecules. The degree of the solvent sorption/desorption is affected by the method of preparation. The nanocomposites prepared by using a solution containing aniline and acetonitrile show a more regular behaviour than those when only pure aniline was in contact with RuCl3. In the presence of an aqueous solution of HCl and aniline a higher degree of aniline incorporation can be achieved. It also attests that the morphology of the layer is influenced by the contacting media. It has also been observed, however, that cycling for a longer period of time in the presence of aniline and HCl causes a decrease in the electrochemical activity of the layer. It is related to the high degree of swelling of the layer, which hinders the charge transfer between the RuCl3 sites. The nanocomposite behaves as a self-doped conducting polymeric system where RuCl x−3 sites act as counterions; however, cation sorption/desorption is inevitable when PANI becomes fully reduced in order to compensate the negative charge of RuCl x−3 . In a smaller extent anion transport also takes place. The self-doped nature of the conducting polymer manifests itself so that PANI remains electroactive even in neutral solutions.