1. Introduction
The steady increase in demand for better and safer batteries for use in electric vehicles, mobile devices, and energy storage has driven research towards the development of new solid polymer electrolytes (SPEs). Among the main properties required for applications of SPEs in devices for technological use are high ionic conductivity, wide thermal and electrochemical window, and stability in different environmental conditions of temperature and humidity [
1,
2]. However, the relatively lower ionic conductivity of conventional liquid electrolytes is the main drawback of SPEs to achieve adequate practical utility.
Among the great variety of SPEs, composites formed by polymers, salts, and nanoparticles have received considerable attention in recent years due to their considerably improved electrical, thermal, mechanical, and optical properties concerning their precursors [
3,
4]. This improvement in the properties of SPEs is obtained thanks to the combination of phases and due to an appropriate interaction between the components. High conductivities in these systems are achieved due to high amorphous phase fractions or by combinations of liquid and solid phases, in which the nanoparticles used as fillers have a high influence on modifying the dynamic conditions and avoiding crystallization.
In particular, carbon nanotubes (CNTs) like single-walled carbon nanotubes (SWCNTs) or multi-walled carbon nanotubes (MWCNTs), which have good mechanical strength, low density, high electrical conductivity, and high thermal stability, can be used to modify the electric and mechanics properties of polymeric matrices [
5,
6,
7,
8,
9].
Although some combinations of polymers with CNTs are used in systems that involve electronic conductivity [
5,
7,
10], many applications in batteries and sensors require ionic conduction. In fact, it was shown that when CNT are added to solid electrolytes they can improve the ionic conductivity of the system [
11,
12]. For instance, some SPEs with lithium ions have shown an increase in ionic conductivity with the incorporation of CNTs, due to intercalation, adsorption, and diffusion of lithium ions into the CNT structure. A high affinity between the electronic cloud of CNTs and Li
+ ions increases the fraction of free ions and provides low lattice energy pathways for ionic migration [
13]. However, there is a possibility that using (CNTs) as fillers in polymers may cause a short circuit between the electrodes of electrochemical devices.
Many different methodologies were used to deposit CNTs and other conductive fillers in polymeric matrices; however, the dispersion of these fillers was a difficult task. For instance, Hammad R. Khaliden and colleagues reported a simple spray dispersion process of CNTs between two polymer layers, reducing the electrical resistance from 12 MΩ to 4 kΩ [
10,
14,
15]. In this work, polymer composites were synthesized, by solution, from the (PEO)
4CF
3COOLi electrolyte added with different amounts of MWCNT, in order to improve ionic conductivity in the electrolyte. The electrical, thermal, and structural properties were studied by impedance spectroscopy, differential scanning calorimetry and infrared spectroscopy. The results indicate that membranes can be used in technological devices such as batteries and gas or moisture sensors.
3. Results and Discussion
Figure 1 shows DSC thermograms between 176 and 238 K for the polymeric solid electrolyte (PEO)
4CF
3COOLi and different combinations of this electrolyte with MWCNT in weight percent, (PEO)
4CF
3COOLi +
x% MWCNT (
x = 0.0%, 0.3%, 0.6%, 2.0%, 4.0%, 8.0%, and 12.0%). The step observed in each case is characteristic of the glass transition temperature (T
g) in these systems [
19,
20]. It is observed that for concentrations up to 2.0% MWCNT, the step assigned to the T
g shifts towards lower temperatures. This shift is related to an increase in the amorphous phase fraction of the new composite. However, for concentrations above 2.0% MWCNT, the T
g shifts towards higher temperatures. Perhaps the increase of MWNCT above 2% might generate more nucleation sites that improve crystallinity [
21]. The T
g values obtained from the DSC thermograms are reported in
Table 4.
DSC thermograms between 302 and 430 K are shown in
Figure 2. The thermogram of a pure PEO membrane shows an endothermic anomaly around 345 K associated with the melting of the crystalline phase of the polymer [
22,
23]. If the polymer is combined with CF
3COOLi, a new endothermic anomaly is observed around 402 K; this anomaly is associated with melting of the complex formed between the polymer and the salt [
24,
25]. When the (PEO)
4CF
3COOLi electrolyte is added with different concentrations of MWCNT, the endothermic anomalies corresponding to the melting of the crystalline phase of PEO and the melting of the complex change in both temperature and enthalpy. This would indicate changes in the crystalline phase fraction of both PEO and the complex formed between PEO and CF
3COOLi.
As seen in
Table 5, when the MWCNT concentration is increased for values up to x = 2.0 wt.%, a decrease in both T
m and ∆H values is observed. For higher concentrations, x > 2.0 wt.%, an increase in T
m values (corresponding to PEO melting, 1st anomaly in
Table 5) is observed. The latter would indicate the formation of MWCNT clusters and polymer chains in the composite and, therefore, an increase in the crystalline phase.
Figure 3 contains the FTIR spectra for PEO, CF
3COOLi, and the different MWCNT concentrations in the solid electrolyte (PEO)
4CF
3COOLi, in the range between 670 and 1310 cm
−1. For CF
3COOLi, the energy band observed at 727 cm
−1 (a in the
Figure 3) is assigned to the CF
3 symmetric bending mode of CF
3COOLi [
3,
26,
27]; this band shifts slightly towards a lower wavenumber (722 cm
−1) when CF
3COOLi is combined with PEO and with the different MWCNT contractions. The peak observed at 797 cm
−1 (b in the
Figure 3) is assigned to the COO scissor vibrational mode [
26,
28].
In the PEO spectrum, the peak observed at 857 cm
−1 (c in the
Figure 3) corresponds to CO stretching vibrations [
11]; these peaks shift slightly to the left in the (PEO)
4CF
3COOLi electrolyte and the different combinations of the electrolyte with MWCNT. The peak observed at 960 cm
−1 (d in the
Figure 3) is assigned to the asymmetric stretching of CH
2 [
11] and remains unchanged in all cases. On the other hand, the peak observed at 1076 cm
−1 (e in the
Figure 3) is assigned to the asymmetric stretching of C-O-C [
11]. However, when PEO is combined with CF
3COOLi and with different MWCNT concentrations, this latter peak shifts to the left, indicating the interaction of Li ions with oxygen atoms of the polymer to form the (PEO)
4CF
3COOLi complex. In turn, the peak corresponding to the CH
2 symmetric stretching vibrational mode is observed at 1242 cm
−1 (h in the
Figure 3) [
11]; these peak shifts slightly to the left (1240 cm
−1) in all systems.
Similarly, the peak observed at 1295 cm
−1 (i in the
Figure 3) in the PEO spectrum is assigned to asymmetric CH
2 torsion [
29]. This peak is found shifted to the left in the electrolyte and the different combinations with MWCNT.
Figure 4 contains the FTIR spectra for PEO, CF
3COOLi, and the different MWCNT concentrations in the solid electrolyte (PEO)
4CF
3COOLi, in the range between 1312 and 3000 cm
−1. The peak observed at 1358 cm
−1 (j in the
Figure 4) is assigned to bending of CH
2 [
29], while the peak observed at 1470 cm
−1 (k in the
Figure 4) is assigned to bending vibration of CH [
29]; the latter remains unchanged in the electrolyte and in the combinations of the electrolyte with MWCNT.
On the other hand, relative to the CF
3COOLi spectrum in
Figure 3, the broad peaks observed at 1146 cm
−1 and 1202 cm
−1 are assigned to the asymmetric and symmetric stretching vibrational modes, respectively (f and g in
Figure 3). In turn, the peak observed at 1673 cm
−1 (l in the
Figure 4) [
17], assigned to the asymmetric stretching of COO, shifts slightly to the right (1686 cm
−1) in the electrolyte and the combinations of this electrolyte with MWCNTs, corroborating the complex formation between PEO and CF
3COOLi.
Figure 5 shows a zoom of the spectrum between 1550 and 1800 cm
−1, where new shoulders are observed at 1655 cm
−1 and 1702 cm
−1 (m and n in the
Figure 5); these are assigned to free ions responsible for the conductivity and ionic aggregates respectively [
17,
30].
Figure 6 shows the Nyquist plots at room temperature (298 K), obtained by impedance spectroscopy for PEO and (PEO)
4CF
3COOLi +
x% MWCNT systems. In the diagrams of PEO and (PEO)
4CF
3COOLi, a straight line can be observed on the right side of the impedance spectrum (low frequencies) followed by a semicircle on the left side of the spectrum (high frequencies). The resistance (R) in these polymeric systems can be obtained by projecting the point where the line meets the semicircle to the Z’ axis. The resistance values show a considerable decrease when PEO and CF
3COOLi are combined, which confirms the formation of the polymer electrolyte (PEO)
4CF
3COOLi and supports the DSC and FTIR results. In the Nyquist plots corresponding to the combinations of the (PEO)
4CF
3COOLi electrolyte with different MWCNT concentrations, the characteristic line of the double-layer capacitance in blocking electrodes is observed, but only in the beginning of the semicircle. In these cases, the resistance is even lower due to the interactions between the electrolyte and the MWNTs.
The values of the real conductivity as a function of frequency were obtained using the relation
where
l is the membrane thickness,
A is the electrode area,
Z′ is the real impedance, and
Z″ is the imaginary impedance.
In
Figure 7, which shows diagrams of real conductivity as a function of frequency (Bode diagrams) for different temperatures, the real conductivity is higher at higher temperatures. This behavior, observed in polymeric solid electrolytes or polymer composites, is due to the agitation of the polymer chains and indicates that ionic conduction in these systems is a thermally activated process. Two regions can be differentiated in these diagrams. One low-frequency region in which an increase of the real conductivity is evidenced. This is due to the polarization of the electrodes by the double-layer capacitance at the interface of the blocking electrodes and the electrolyte. In a second region, where the real conductivity remains constant as the frequency increases (plateau), behavior is associated with the movement of long-range ions (hopping process) [
31]. The conductivity value in the latter region is assumed to be the value of the dc membrane conductivity.
Figure 8 shows plots of the dc conductivity as a function of the inverse of temperature for different MWCNT concentrations in the (PEO)
4CF
3COOLi electrolyte. For all systems, it is observed that the dc conductivity increases with increasing temperature following the Vogel Tamman Fulcher (VTF) equation [
18,
32]:
where
σdc is the conductivity obtained in the plateau,
σ0 is the pre-exponential factor,
EA is a parameter known as pseudo-activation energy,
kB is the Boltzman constant,
T is the absolute temperature, and
T0 stands for the temperature (below the glass transition temperature) at which the free volume, responsible for ionic conduction, disappears, and the ionic conductivity cancels out.
Table 6 shows the parameters for the fitting to Equation (2).