Introduction

Lithium-based batteries have become fundamental storage devices for a wide variety of applications ranging from green transportation to portable electronic devices. The performance of the device strictly depends on the materials used. In particular, improved nanostructured electrodes based on high-energy-density materials and novel polymer electrolytes able to reliably replace the organic carbonate-based ones are the main research fields in the present years.

Solid polymer electrolytes (SPEs) are electrolyte having a polymer as host matrix [1–3]. SPEs are free standing and dimensionally stable film and do not contain any solvent. The pioneering work on SPEs was initially done to overcome the safety problem with regard to liquid electrolytes. However, SPEs generally suffer of low ionic conductivity. The previous works in this field were reported by using high-molecular-weight poly(ethylene oxide) (PEO) as polymer host [4–6]. Due to well-known crystalline properties of PEO, it was conjoined with various additives to soften the polymer backbone in order to increase the ionic conductivity [7, 8]. Moreover, in order to facilitate its upscalability and lower its cost, UV-induced photopolymerization processes [9] have been developed for the preparation of SPEs without requiring the use of solvents and catalysts.

Oxygen atoms in polymer host are known to be the polar active site for lithium ions to interact, form complexes, and therefore enhance the ionic conductivity value of SPEs [10–12]. Previous studies reported that Li+ tends to interact with the oxygen atoms, thus indirectly be transported through the segmental motion of polymer chains [13, 14]. In addition, the lone pairs that were carried by those polar functional groups were believed to contribute to the effective salt solvation [10, 11, 14]. This behavior attracted much attention from the scientific community to focus on a various type of polymer host containing many polar atoms [15, 16]. Methacrylate-based materials become one of the predominant polymer hosts for many works, due to its low processing cost and high chemical and mechanical stability [10, 13]. In fact, poly(glycidyl methacrylate) (PGMA) possesses good mechanical properties, is cheap, and can be easily processed through UV curing method [17–19]. This material contains several oxygen-based functional groups, such as C=O, C–O–C, and epoxy ring. Thus, the use of PGMA as polymer host is useful to provide well-solvated salt through its polar functional group, also helping the transport of Li+ ions through its segmental motion.

The development of good SPEs is not only depending on the polymer host but also concerning with the salts used for SPE preparation. Several works on SPEs using different type of polymer hosts showed significant increment in ionic conductivity after doping the polymer matrix with salts. On the other hand, the decreasing trend of ionic conductivity after the addition of salt between 20 wt.% and 30 wt.%. was also reported [10, 20, 21].

Despite the extensive attention to SPEs, the mechanism by which ions are transported through the polymer films is poorly understood. The choice of lithium bis(trifluoromethylsulfonyl)imide (LiTFSI) salt presented in this study provides a further tool for the comprehension of the role of this kind of salt, constituted of big anions introduced in the PGMA matrix. To the best of our knowledge, PGMA doped with LiTFSI as the charge carrier has never been reported yet, but our previous results with PGMA doping with iodide salts for photoelectrochemical cells [18, 19] make us believe that this materials could be of great interest in the field of lithium batteries [22]. Thus, this paper focuses on the effect of LiTFSI on the chemical interaction, structural, thermal properties, ionic conductivity, and lithium transference number of PGMA–LiTFSI SPEs.

Experimental

Materials

Glycidyl methacrylate (GMA), 2,2-dimethoxy-2-phenylacetophenone (DMPP), and lithium bis(trifluoromethylsulfonyl)imide) (LiTFSI) were purchased from Sigma-Aldrich. Ethanol (95 % purity) was provided by Systerm, and tetrahydrofuran (THF) was purchased from J.T. Baker. All chemicals were used without further purification.

Sample preparation

PGMA were prepared by exposing an appropriate amount of GMA and DMPP solution under UV lamp (Rs Ltd) and nitrogen atmosphere for 18 min. This method was modified from the report of Hanifah et al. [19]. The resulting PGMA was then washed with ethanol to dissolve unreacted monomer. PGMA was used to prepare SPE PGMA–LiTFSI at different weight percentages of LiTFSI by solution casting technique. Both PGMA and LiTFSI were dissolved in THF and stirred until a homogenous solution was obtained and casted into a Teflon mould. THF was then dried at 30 °C in vacuum oven for 24 h.

Characterization

FTIR analysis was carried out using PerkinElmer Spectrum with frequency ranging from 4000 to 650 cm−1, with a scan resolution of 4 cm−1. X-ray diffraction pattern of polymeric films was performed by using Bruker D8 Advance diffractometer. Data were collected from the range of diffraction angle 2θ of 10° to 40°. Glass transition temperature was measured using Mettler Toledo differential scanning calorimeter under nitrogen, with a scanning rate of 10 °C min−1. Ionic conductivity was measured by HIOKI 3532-50 LCR HiTESTER. The polymer film was sandwiched between stainless steel electrodes, and a frequency range of 50–5 MHz was adopted at ambient temperature. Lithium transference number measurements were carried out using Zive multichannel potentiostat/galvanostat/FRA. The polymer film was sandwiched between lithium electrodes and were measured through DC polarization technique at ambient temperature, and the data were collected every 10 s for 2.5 h. The potential difference applied was equal to 10 mV.

Results and discussion

SPEs were prepared by a fast, cheap, and upscalable process of UV-induced polymerization. FTIR analysis was carried out on photopolymerized samples, and Fig. 1 shows the region of carbonyl functional group in different PGMA–LiTFSI mixtures. The absorption of carbonyl functional group for PGMA is observed at 1723 cm−1 for polymer electrolytes at all concentration values of LiTFSI. However, it is interesting to notice that with the addition of LiTFSI at different loadings, the intensities and the sharpness of C=O peak were gradually reduced. Since the intensity of an absorption band is related to the change in dipole moment and to the number of coordination, it seems that the reduction of the C=O peak intensity is due to the increased ion amount. The most probable reason for this behavior is the presence of interactions between Li+ ions with the partially negative charges of oxygen atoms in the carbonyl functional groups. The same result also was reported in the recent publication of polymer host that contains carbonyl functional groups and lithium salt [10, 14].

Fig. 1
figure 1

FTIR spectra in the carbonyl group region for UV-cured PGMA-based SPEs

Full size image

Figure 2 shows asymmetric stretching mode of ether functional group for PGMA, which is observed at 1146 cm−1. Peak at 1253 cm−1 for PGMA corresponds to weak ring stretching band (breathing mode) of epoxy, and it is shifted to 1257 cm−1 with lower intensity when a 40 wt.% of LiTFSI is adopted. Figure 3 shows that there are two strong ring deformation bands for epoxy ring at 905 cm−1 (asymmetric) and 844 cm−1 (symmetric) for both ether and epoxy units, respectively. The symmetric stretch of the latter was slightly shifted to 848 cm−1 in the presence of 40 wt.% LiTFSI. According to Uma et al., coordinations that occur between polymer host and salt often induce changes in the vibrational mode of the functional groups [23]. However, there are no significant changes or shifting of the stretching frequency in the present PGMA-based samples.

Fig. 2
figure 2

FTIR spectra in the 1000–1400-cm−1 region for UV-cured PGMA-based SPEs

Full size image
Fig. 3
figure 3

FTIR spectra in the 650–1000-cm−1 region for UV-cured PGMA-based SPEs

Full size image

As regard the TFSI− anions, FTIR spectra showed two asymmetric stretch (1346 and 1326 cm−1) and symmetric stretch (1134 cm−1) of SO2, strong SO stretch (1060 cm−1), and multiple CF band (1243, 1190, 800, and 748 cm−1). From Fig. 2, SO2 and SO peak in LiTFSI shifted to lower wave number of 1131 and 1056 cm−1 when 40 wt.% LiTFSI was used. Besides, CF peak in LiTFSI salt shifts to 1229, 1183, 790, and 743 cm−1 at 40 wt.% LiTFSI concentration. These significant changes occurred due to the lone pair interactions of oxygen and fluorine atoms with Li+. Therefore, the results contribute to bond weakening reducing the stretching frequency [24]. According to Borodin et al., there is high probability of Li+ to coordinate with polar atoms in TFSI− at high ion concentration [25]. However, the suggestion is that fluorine atoms will form coordination bond with Li+ due to its higher electronegativity compared to that of oxygen atoms and also assist the oxygen atoms in TFSI− anions to form further coordination. Therefore, huge formation of ion aggregates is believed to occur at high salt concentration, which indicates very poor ion solvation by the polymer chains. The discussed atoms in the structure of PGMA and LiTFSI salt are shown in Fig. 4.

Fig. 4
figure 4

PGMA and LiTFSI salt molecular structure

Full size image

SPE-based PGMA–LiTFSI diffractograms are shown in Fig. 5. The sharp peak shows crystalline region while the broad peak represents the amorphous region of the polymer matrix. As shown in Fig. 5 (b–f), there are prominent semicrystalline curves of pure PGMA at 2θ = 15°–22° in all diffraction patterns of SPEs. However, there are continuous reductions of the peak intensity upon the addition of salt. The presence and the interaction of ions in the SPEs suggest the disruption of the rigid segment arrangement of polymer chains [20, 26]. Based on the previous reports, amorphous region in SPEs contributed in the ionic conductivity enhancement [4, 6, 27]. Figure 5 (a) represents the LiTFSI salt diffractogram: very sharp peaks are obviously present. However, no salt peaks appeared even at 40 wt.% LiTFSI-laden PGMA, thus highlighting that no salt association occurred in the SPE. This is a positive fact, because it ensures the good salt dispersion in the polymer matrix and also its availability when employed in an electrochemical cell.

Fig. 5
figure 5

XRD diffractogram of LiTFSI salt (a), PGMA (b), 10 wt.% LiTFSI in PGMA (c), 20 wt.% LiTFSI in PGMA (d), 30 wt.% LiTFSI in PGMA (e), and 40 wt.% LiTFSI in PGMA (f)

Full size image

Glass transition temperature (T g) values of PGMA and PGMA–LiTFSI SPEs are listed in Table 1. T g value was determined using mid-point method on DSC curve. For PGMA, the T g is 318 K and the value shifted downward in SPEs. As the T g of a polymer is the temperature at which the segmental motion begins, this result confirms that very small or no significant interactions occur between Li+ and oxygen atom in the polymer host [28]. As a result, the T g decreases due to the increase interchain voids.

Table 1 Ionic conductivity (σ) and glass transition temperature values (T g) of PGMA–LiTFSI SPEs
Full size table

Table 1 shows the room temperature ionic conductivity value for PGMA and SPEs at different weight percents of LiTFSI. The conductivity of PGMA is 3.09 × 10−10 S cm−1. Ionic conductivity of PGMA–LiTFSI increases with the increasing weight percentage of LiTFSI in the system. The highest ionic conductivity obtained is at 40 wt.% LiTFSI and is 4.96 × 10−8 S cm−1. The overall increment in ionic conductivity is due to the increment of the number of charge carriers in the system [21, 29]. There is a pronounced increase of ionic conductivity from 10 wt.% up to 30 wt.% LiTFSI salt. However, low increment of ionic conductivity value was observed from 30 wt.% up to 40 wt.% of LiTFSI. At low LiTFSI salt concentration (10 wt.% and 20 wt.% LiTFSI), the lack of charge carriers led to low ionic conductivity values, while when the salt amount is too high, ion aggregation effect overcomes the effect of an increased charge carrier number [30]. The obtained ionic conductivity value (4.96 × 10−8 S cm−1) should be contextualized to state of the art. Shim et al. measured conductivity values around 10−5 S cm−1 by using core–shell silica particles with ion-conducting poly(ethylene glycol) and anion-trapping boron moiety in the shell layer [31]. Young et al. investigated polypropylene–poly(ethylene oxide)–polypropylene triblock copolymers as SPEs for lithium batteries and measured ionic conductivity values equal to 10−5 S cm−1 at room temperature [32]. A similar value was also measured by Han et al. with a crosslinked SPE fabricated by a solvent-free hot-pressing method adopting an ethylene oxide (EO)/LiN(SO2CF3)2 (LiTFSI) molar ratio of 20:1 [33]. At first glance, the ionic conductivity reported in this study is lower than values reported by other groups. However, it is necessary to consider that the polymer electrolyte described in this work was obtained by a simple photopolymerization step, i.e., without the use of catalysts, thermal ramps, and processes of separation/purification. In view of a large-scale diffusion of solid-state batteries, this material is very promising in terms of both cost and preparation time.

Some of the promising SPEs for lithium batteries possess wide temperature range of high ionic conductivity. Thus, also in our case, it is necessary to evaluate the SPE conductivity dependence upon temperature. Therefore, the ionic conductivity of SPEs containing different salt amounts was measured from 40 to 100 °C. Experimental data deviated from Arrhenius equation. However, they obeyed Vogel–Tamman–Fulcher (VTF) equation:

$$ \sigma =\frac{A}{T^{\mathrm{m}}}{\mathrm{e}}^{-\frac{E_{\mathrm{a}}}{T-{T}_0}} $$
(1)

where σ is the ionic conductivity, A is the pre-exponential factor, E a is the apparent activation energy for the charge transport within SPEs, T 0 is the 30 K/50 K less than glass transition temperature value of SPE, T is the temperature at which the ionic conductivity is measured, and m is a factor which is ½ [34, 35]. We considered a T 0 value 50 K less than the experimental T g value of SPEs, which gave us the best fit of Eq. 1 to the experimental data. From Table 2, the activation energy values of SPE at 10 wt.%, 20 wt.%, 30 wt.%, and 40 wt.% LiTFSI are 0.070, 0.068, 0.054, and 0.067 eV, respectively. The reducing trend of activation energy from 10 wt.% LiTFSI until 30 wt.% LiTFSI is shown in Fig. 7. Among all of the concentrations, SPE at 30 wt.% LiTFSI exhibited the lowest activation energy, which was 0.054 eV. At low LiTFSI concentration, the activation energy of SPE is high because of the higher crystallinity of the SPE system. Furthermore, the activation energy is increased at 40 wt.% LiTFSI. This is inconsistent with the increase of amorphous regions in the system, as indicated by the X-ray diffraction (XRD) diffractogram and by the lowest glass transition temperature of this SPE. The higher E a value for SPE at 40 wt.% LiTFSI indicated the higher energy needed for the Li+ to mobile. This result is in agreement to the first expectation that at high salt concentration, the probability of ion triplets or aggregations is high and this reduced Li+ mobility, therefore adversely affecting the ionic conductivity. According to Olsen and Uma, pre-exponential factor, A, is related to the number of charge carriers and mobility of the ions in electrolyte system [23, 31]. From the results obtained, it is found that the pre-exponential values are the same for 30 wt.% and 40 wt.% LiTFSI amount. This confirm that, at 40 wt.%, there is limited mobility of Li+ in the system which contributes to slow ionic conductivity enhancement. In Fig. 6, the ionic conductivity value was increased with the increased temperature, due to the ability of Li+ to dissociate from its counter ions and be more mobile.

Table 2 Pre-exponential factor, apparent activation energy, and glass transition temperature value
Full size table
Fig. 6
figure 6

Vogel–Tamman–Fulcher plot for SPEs

Full size image

Lithium ion transference number is one of the methods that could be useful to determine the cationic transportation within the polymer host. It is the fraction of the total current that is carried by a specific ion in a medium [36]. As it has been reported by Hiller et al. [37], the T Li+ can be calculated by ignoring the interfacial resistance if the potential difference is between 10 and 20 mV. The transference numbers were calculated with the following equation:

$$ {T}_{{\mathrm{Li}}^{+}}={I}_{\mathrm{T}}-\frac{I_{\mathrm{e}}}{I_{\mathrm{T}}} $$
(2)

where I T is the total current and I e is the residual current [33]. SPE at 30 wt.% LiTFSI was chosen to undergo the lithium transference number measurement due to its lowest activation energy, as shown in Fig. 7. Based on Fig. 8, the calculated TLi+ of SPE at 30 wt.% LiTFSI is 0.98. This has proved that lithium ions dominate as the charge carrier in this system.

Fig. 7
figure 7

Activation energy for SPEs at different LiTFSI weight percentages

Full size image
Fig. 8
figure 8

Plot of current versus time of SPE at 30 wt.% LiTFSI

Full size image

Conclusion

The results obtained showed that the proposed series of free-standing PGMA–LiTFSI SPEs films was successfully prepared and characterized. FTIR spectra analysis confirmed that interaction occurred at oxygen and fluorine atoms. Lithium ions form coordination with polar atoms in TFSI− units and limit the number of charge carrier and ion mobility. The increase of amorphous region indicated by XRD diffractogram proved that the presence of LiTFSI ions has weakened the rigid polymer structure; the glass transition temperature decreased for the same reason. Temperature-dependent measurements showed that the SPE system obeyed VTF equation during conductivity analysis. The present findings highlight that salt concentration and the type of anion play an important role in studying the ion transport in SPE system. We intend to expand our studies’ scope by adding ionic liquids into these SPE systems in order to further understand how anions influence the transport of lithium ions in polymeric matrices, to be then applied in Li-ion batteries.