Among advanced energy storage systems, electrochemical capacitors or supercapacitors (SC) find increasing application, their key advantages being high power density allowing rapid accumulation and energy release, long service life, and high performance over a wide temperature range, including extremely low temperatures [1, 2]. Due to their good performance, SC have found application in alternative power devices, including independent power sources for use in Arctic regions, as backup power sources for essential consumers, in electric vehicle braking energy recuperation systems, and in hybrid vehicles [3–6]. The use of SC is of special importance for peak load conditions, and therefore their operation in combination with batteries and fuel cells is a critical requirement, allowing one to increase the stability of energy systems, reduce network loads, and provide for the integration of renewable energy sources into power supply systems [7, 8].

In recent years, the attention of researchers has been focused on the development of energy accumulation systems for portable electronics devices, i.e., portable, lightweight, and comfortable accessories which people wear and use for intellectual assistance, navigation, communication, health monitoring, and other purposes. Such devices can be worn outside such as rings, badges, watches, glasses, bijouterie, shoes, or clothes, or inside in implants, such as cardiac pacemakers or neuroprosthetics [4, 6, 9–11]. Special attention is paid to portable devices used for health monitoring, testing, and recording of metabolic changes in human body, i.e., blood sugar, temperature, arterial pressure, heart rate, health condition during sleep etc., which can be vitally important for a large number of people. Great possibilities are expected to emerge in the field of remote health monitoring due to the evolvement of the so-called Internet of Things, i.e., systems of interconnected devices which can automatically collect and transfer data through wireless networks. It is believed that the Internet of Things will play a decisive role in supporting stable health parameters of patients in hospitals and in their everyday life. Due to the permanently increasing energy consumption by portable devices, there is the need for self-charging power sources which could use the energy of light, mechanical motion, human body heat etc. for recharging by converting them to electric power. A separate and quite actual trend of this research is the development of power sources for bionic robotic prosthetics and neuroprosthetics [4, 9, 12, 13]. Energy accumulation systems which are unalienable part of such power sources should meet a number of requirements: along with the main criteria related to energy and power performance, there are additional requirements, including non-toxicity and fire safety, mechanical impact stability, high temperature tolerance, pulsed mode operation capability, charging mode variability, and flexibility. It was shown that SC meet the above requirements the most completely, and furthermore exhibit a lower tendency to the degradation of energy and power performance in miniaturized options, in comparison with lithium-ion batteries [3, 9, 10]. It should be stressed that for this application, of key importance is the capability to give back and receive current pulses, which is required for data collection and transfer and for the mechanical motion of robotic prosthetics, rather than the power capacity of the storage system.

The wide practical implementation of SC is still hindered by their relatively low specific power capacity as compared to batteries. Therefore it is an important task to develop SC and micro-SC having the highest specific power capacity for this range of devices (about 10–20 (W∙h)/kg) and also meeting the above requirements. There are different approaches to increasing the power capacity of SC. In conventional symmetrical two-layered SC consisting of two porous conducting electrodes separated by an electrolyte-impregnated spacer, energy is accumulated in the form of double electric layer (DEL) charge at the electrode/electrolyte boundary. Since achieving the highest specific capacity requires the greatest possible contact area between the electrolyte and the electrode, the latter are typically made from nanoporous carbon materials with a specific surface area of above 1000 m²/g. The development of carbon materials with optimized porous structure hierarchy and the use of new carbon materials such as carbon nanotubes, including those doped with heteroatoms, graphene-based electrode materials, and activated carbon fiber and tissues allows one to achieve a specific capacity of up to 200 F/g for water-based electrolytes [14–17]. Another approach is to modify the electrolyte by adding so-called pseudo-capacity materials capable of rapid reversible oxidation-reduction processes during SC charging and discharging, transition metal oxides, electrically conducting polymers, and metal-organic skeleton structures with embedded chemically active and functional elements. To increase the electrical capacity, one can also use asymmetrical structures in which the positive and negative electrodes and made from different materials. This allows one not only to make the pseudo-capacity to contribute to the overall capacity of SC but also to increase its operation voltage, thus significantly increasing the charge density and power. Theoretically, pseudo-capacity electrode materials can allow one to increase the power capacity of SC by several times, but in practice one can rarely achieve a more than two-fold increase [18–22].

Taking into account that the specific energy of SC increases as the square of the difference of potentials between the electrodes, a promising strategy of increasing the power performance of SC is to increase the operation voltage of the cell, which is mainly limited by the electrochemical stability of the electrolyte. Whereas for water-based electrolytes (salts, alkalis and acid solutions) the electrochemical window is 1.23 V, that of organic electrolytes and ionic liquids can be increased to 2.7–3.0 V, and therefore most SC manufacturers use organic electrolytes, most typically acetonitrile-based ones (AN). AN-based electrolytes have the highest electrical conductivity among organic electrolytes, thus providing for sufficiently high power performance and efficiency of high-power SC, i.e., up to 10 kW/kg and 95–98 %, respectively [2, 23]. However, the use of electrolytes on the basis of AN or other nitriles requires taking into account safety aspects, since these materials are noxious, volatile, and inflammable liquids with high vapor pressures and low flash points. Incidental damage or overheating of AN-based electrolyte devices to above 60 °C, e.g. by direct sunlight, can cause SC depressurization due to high vapor pressure, AN vapor emission, and flashing, resulting in a release of extremely noxious products.

Considering the above requirements to SC electrolyte for use in portable devices, one should bear in mind that the electrolyte should deliver the highest possible capacity and energy parameters of SC, along with a number of performance characteristics, i.e., sufficiently wide range of operation temperatures combined with minimum toxicity and maximum fire safety. Generalized relationship between the required SC electrolyte properties and the physicochemical parameters of its components that are required for the achievement of the above properties is shown in Fig. 1. Using the data presented in the illustration, one can justify the choice of solvents for SC electrolytes and set out requirements to the properties of the ionogenic salt used.

To provide for the highest electrochemical stability and electrical conductivity, the electrolyte solvent should be an aprotonic high-polarity liquid with a sufficiently wide liquid phase existence temperature range he. The solvent should be chosen from liquids having high dielectric permeability providing for sufficient solubility and degree of dissociation of ionogenic salts into ions, as well as low viscosity for achieving the highest electrical conductivity of the electrolyte. Taking into account that the minimum viscosity is achieved if the sizes of molecules are small, the range of solvents complying with the above requirements is limited to nitriles and cyclic alkylcarbonates with sufficiently low molecular weights [23, 24]. From the viewpoint of physicochemical parameters, nitriles (AN, propylnitrile, and metoxyacetonitrile (Table 1)) meet the above requirements the most completely. Those solvents are quite rarely used in electrolytes of lithium-ion batteries due to their insufficient stability against reduction, but they are the optimum solvents for high-power SC electrolytes.

However, taking into account SC operation safety requirements for portable devices, a viable alternative for nitriles can be another group of high-molecular organic solvents, i.e., cyclic alkylcarbonates, such as ethylenecarbonate (EC), PC, butylenecarbonate, vinylenecarbonate (VC), and fluoroethylenecarbonate (FEC), Table 1.

Alkylcarbonates are non-toxic and have low vapor pressure and high boiling and flash points. In SC, alkylcarbonate-based electrolytes are mainly contained in spacer and electrode pores, and electrolyte leakage will not occur even if the container is damaged.

However, the uniquely high dielectric permeability of alkylcarbonates is combined with their high viscosity and melting point, and therefore they are rarely used alone as electrolyte solvents. The best candidate is PC which has moderate viscosity and the widest liquid phase existence temperature range. Note that PC is used quite rarely in lithium-ion battery electrolytes with graphite anodes because it can cause graphite corrosion. Most typically, lithium-ion battery electrolytes are made from EC mixed with co-solvents, and VC and FEC are used as additions favoring the formation of an interphase contact layer.

According to literary data [24], direct replacement of AN for PC in electrolyte with TEA∙BF₄ ionogen leads to an almost five-fold decrease in the specific electrical conductivity due to the high density of PC (Table 1), which, in turn, has a negative effect on the power performance of SC. This drawback can be corrected by optimizing the electrolyte components and their concentrations by selecting an alternative ionogenic salt, choosing optimum component concentrations, and introducing additions to electrolyte.

Regarding ionogen choice, the primary requirement to this salt is good solubility in PC (at least 1.5–2 mole/l at 25 °C), high electrochemical stability, and stability against moisture. Inorganic salts having the required solubility in PC are lithium and sodium hexafluorophosphates, and among organic ones, such are alkylammonium and alkylphosphonium tetrafluoroborates and hexafluorophosphates. According to literary data [24–26], the most electrochemically and hydrolytically stable materials are alkylammonium tetrafluoroborates. Furthermore, the electrical conductivity of their solutions in PC is higher than those of inorganic salts.

For the formation of DEL having the highest capacity, ionogen cations and anions should have the smallest possible size. At the same time, reaching sufficient salt solubility in organic solvents, high dissociation degree, and high electrical conductivity requires choosing ionogens having lower crystalline lattice energy and hence large ions. For example, tetramethylammonium tetrafluoroborate that has a smaller cation has a solubility of within 0.1 mole/l and therefore cannot be used for solutions with sufficient electrical conductivity. Analysis [25, 26] suggests that the best ionogens can be tetraethylammonium tetrafluoroborates which contain 6–8 carbon atoms in the cation.

The most commercially available ionogen which is often used in AN-based electrolytes is TEA·TFB. However, the solubility of that salt in PC is limited, even at room temperature (Table 2), which is insufficient taking into account that the optimum ionogen concentration in electrolyte should be 1.0–1.5 mole/l. With a decrease in temperature, electrolyte components can start to crystallize in electrode pores and on its surface.

In this work, four ionogenic salts were studied: TEA·TFB, TEMA·TFB, SBP·TFB, and DMP·TFB, of which two contain anticyclic cations and the other two are formed by cyclic cations (see Table 2). All the test salts have high solubilities in PC. Furthermore, there are indications that the cyclic structure of the SBP·TFB cation can favor higher electrochemical stability of electrolyte. In recent years, data have been published on the use of the DMP·TFB salt, which also contains a five-member cycle and has the smallest cation size [26–30].

Although PC-based electrolytes have already been used in a number of commercial SC, the available literary data on their capacity and performance, e.g. maximum operation voltage, permissible operation temperature range, and durability are insufficient for a comprehensive evaluation of their potential practical applications in the field considered. Furthermore, SC parameters depend largely on the type of ions constituting ionogenic salts, because they participate directly in DEL formation, but there are only scarce literary data on the possibility of using other salts than TEA∙BF₄ for SC electrolytes [31, 32].

The aim of this work was to systematically study the electrochemical properties and performance of PC- and alkylammonium tetrafluoroborate-based SC electrolytes containing both non-cyclic and cyclic cations, i.e., TEA·TFB, TEMA·TFB, SBP·TFB, and DMP·TFB, for understanding their effect on the properties of electrolytes and choosing an ionogenic salt which can deliver the optimum combination of SC parameters for portable electronics applications.

The first stage of our study was aimed at understanding the effect of ionogenic salt type on the specific capacity and specific energy of SC. To this end, laboratory SC cells with 1 M solutions of SBP∙TFB, DMP∙TFB, TEA∙TFB, and TEMA∙TFB salts in PC as electrolytes were studied using the GCD and CVA methods (Figs 2a and 2b). The shapes of the GCD curves are close to equilateral triangles, suggesting the implementation of a double-layer energy accumulation mechanism (Fig. 2a). The CVA curves are also almost perfect close-to-trapezoidal, typical for double-layer SC (Fig. 2b). The specific energy and capacity of the SC cells with electrolytes containing the test ionogens, as calculated on the basis of the GCD data, are generally close, suggesting quite minor differences between the test ionogens from the viewpoint of their capability to form DEL (Table 3). The TEA∙TFB salt electrolytes have slightly lower capacity due to a large size of the salt cations.

The next stage of the study was the estimation of the electrochemical stability window for SC cells with the test electrolytes at 25, 55, and 85 °C. Based on the CVA curves measured in a voltage range of 2.5 to 4 V (Fig. 3), the electrochemical stability window for all the test electrolytes at 25 °C can be estimated to be 3.5 V. With an increase in the operation voltage to above that point, redox processes of electrolyte components are initiated, which can be seen in the CVA curves as an abrupt increase in the current density.

At 55 °C the electrochemical stability window for SC with the test electrolytes is narrower (Fig. 4). Whereas at room temperature there is almost no difference between the electrochemical stability of the electrolytes, an increase in the temperature to 55 °C and, especially, to 85 °C clearly demonstrates the advantages of the ionogens containing cyclic cations, i.e., SBP∙TFB and DMP∙TFB (Table. 4). The CVA data suggest that, in the whole test temperature range, the electrochemical stability window is the widest for the SC with the SBP∙TFB ionogen (3.3 V); for the DMP∙TFB electrolyte, the 3.5 V operation voltage range is retained at 55 °C but decreases to 3.0 V at 85 °C. Thus, according to the CVA data, the highest electrochemical stability over the whole test temperature range is observed for the ionogens with cyclic cations: SBP∙TFB and DMP∙TFB.

From the viewpoint of practical applications, one of the most important SC parameters is high durability, evaluated as the number of charging/discharging cycles after which the SC still has acceptable performance. As a rule, double-layered SC can withstand more than 500,000 charging/discharging cycles without specific capacity reduction by more than 20 %. Durability tests under close-to-real SC operation conditions require very long time, and hence the stability of SC cells and their components is tested using various accelerated durability testing methods imitating SC operation in different modes. One of such testing methods, which characterizes the degradation susceptibility of electrolyte components during storage in a charged condition, implies SC cell exposure to DC voltage for a long time with periodical capacity measurements. For this test, SC cells with the test electrolytes were exposed for 2 h at 55 °C and then for 2 h at 85 °C at a 1.35 V DC voltage, with CVA curves in a 0 to –3 V range being recorded after each stage (Fig. 5). It can be seen from Fig. 5 that after 55 °C exposure, cells with all the test electrolytes demonstrate almost identical electrochemical behavior. At 85 °С, the situation changes for SC cells with the TEA·TFB and TEMA·TFB salts: they exhibit a decrease in the current density, and hence the capacity, over the whole voltage range, most probably due to the electrochemical decomposition of ionogens at elevated temperatures. For SC cells with the SBP·TFB and DMP·TFB salts, 85 °C exposure reduces the capacity but slightly.

After the latter stage of the study, based on the overall scope of experimental data, more specifically, due to its low specific capacity and energy, as well as low thermal stability, TEA∙TFB salt electrolyte was discontinued from further study as the least promising one.

Further durability tests of SC cells were carried out in continuous GCD mode in a voltage range of 1.35–2.7 V at 25 to 90 °C. The cells were preliminarily cycled for a certain time at room temperature, and then GCD tests were run for 1000 cycles at 60, 80 and 90 °C each. After 1000 cycles at the above temperatures, the cells were cooled down to room temperature and studied using the CVA method. This allowed us to identify a specific temperature range for each electrolyte in which intense degradation processes onset. The experimental data confirm the conclusions made on the basis of the previous test series. The TEMA∙TFB salt electrolyte lost performance even before the completion of the 80 °C tests, which was indicated by significant changes in its CVA curves after cycling at that temperature (Fig. 6a). At 90 °C the cells filled with that electrolyte did not work. The capacity of the SC cells with SBP∙TFB salt electrolyte declined after each cycling stage (Fig. 6b), but they retained some performance even at 90 °C. Finally, the DMP∙BF₄ salt electrolyte proved to be electrochemically stable at 80 °C (Fig. 6c), its CVA curves changing but slightly after cycling at that temperature. A decrease in the capacity and an increase in the resistivity were only observed after cycling at 90 °C. Thus, the DMP∙BF₄ + PC electrolyte proved to be the most electrochemically stable over the whole test temperature range.

Finally, the SC cells with SBP∙TFB+PC and DMP∙TFB + PC electrolytes, which demonstrated the highest thermal and electrochemical stability during the previous test stages, were exposed to long-term cycling in the 50 to 95 °C range in continuous charging/discharging mode at 1.35–2.7 V for 10,000 cycles at each test temperature. The SC specific capacity vs number of cycles curves for different temperatures (Fig. 7) prove the high stability of the DMP∙TFB + PC electrolyte: SC with this electrolyte retained 80 % of the initial capacity after 50,000 charging/discharging cycles at 50 to 85 ℃ and 77 % of the initial capacity after 70,000 cycles at 50 to 95 ℃. Under these conditions, the SC with the SBP∙TFB + PC electrolyte behaved in a less stable manner: one of the cells lost 20 % of capacity at as low as 50 °C, and the other cell, at 80 °C. The capacity vs number of charging/discharging cycles curve has a “saw-like” pattern which is typically attributed to electrolyte components degradation processes accompanied by gas release.

Thus, the durability tests confirm the electrochemical stability of the DMP∙TFB + PC electrolyte at up to 95 °C. At up to 85 °C, SC with that electrolyte can be operated at 3 V, indicating a high specific energy reaching 30 (W·h)/kg. The choice of ionogen is a key factor in increasing the durability of SC. Alkylammonium tetrafluoroborates with cyclic cations, especially DMP·TFB, demonstrate high electrochemical stability at elevated temperatures, which is unachievable with conventional TEA·TFB and TEMA·TFB. It should stressed that the possibility of SC operation at such elevated temperatures is one of the key advantages of the PC-based electrolytes over their AN-based counterparts which are not recommended for operation at above 60 °C.

In order to find the lower limit of the DMP∙TFB + PC electrolyte SC operation temperature range, we measured the capacity of the cells at low temperatures, down to –40 ℃. It can be seen from Fig. 8 that the SC capacity decreases significantly with a decrease in temperature. At 0 °C, the cell capacity is 80 % of the 25 °C figure, at –20 °C the capacity decreases twofold as compared to the room-temperature one, and at –40 °C the SC with that electrolyte loses 83 % of the initial capacity, which is accompanied by an increase in its equivalent serial resistivity by 360 % (Fig. 8). The high resistivity at low temperatures is caused by an increase in the electrolyte viscosity which abruptly decreases the ion migration rate and hinders the formation of DEL. Based on the criterion according to which SC should retain at least 80 % of its rated capacity, one can conclude that the DMP∙TFB + PC electrolyte has an operation temperature range of 0 to 85 °C.

The study showed that the PC-based electrolyte with DMP∙TFB ionogen containing cyclic cation is the most suitable from the viewpoint of requirements to electrolytes of energy storage devices used in portable electronics. Due to its good electrochemical stability, this electrolyte provides for an operation voltage of 3.0 V thus achieving a specific energy of about 30 (W·h)/kg.

As compared with its counterparts, DMP∙TFB has the highest electrochemical stability at elevated temperatures of up to 85 °C, which was experimentally confirmed by continuous GCD tests for 70,000 cycles. The recommended operation temperature range for SC with DMP·TFB+PC electrolyte at 3.0 C is 0–85 ℃, with the possibility of short-term temperature growth to 95 °C.

Further studies should be carried out aimed at developing milticomponent co-solvent systems which would increase the electrical conductivity of PC-based electrolytes over the whole operation temperature range and improve their negative-temperature performance, as well as exploring the possibility of using new-generation electrode materials in combination with the electrolytes developed herein.