**4. Discussion**

The investigation and individuation of new solvents is a relevant issue that has wide implications in different industrial sectors, and, among them, energy storage represents a driving force for their development. In the field of supercapacitors, the most common solvents used for the electrolytes are acetonitrile (ACN) and propylene carbonate (PC).

The use of ACN has several contraindications due to its high vapor pressure (9.71 kPa at 20 ◦C) and low flash and boiling points. Indeed, according to Regulation (EC) No. 1272/2008 (Classification, Labeling and Packaging (CLP)) it is classified as a "highly flammable liquid and vapor", and as a volatile organic compound (VOC). Moreover, ACN exhibits acute toxicity for organs and tissues through different types of exposure (skin contact, ingestion, inhalation).

PC is a more user-friendly solvent compared to ACN: it has low vapor pressure (0.006 kPa at 25 ◦C), high flash and boiling points, and it is classified without hazard statements relative to inhalation or skin exposures. However, the toxicity of organic cyclic carbonates is still under investigation. In recent studies, Strehlau et al. [58] reported that these compounds can penetrate in vitro the simulated blood-cerebrospinal fluid barrier, and it is therefore assumed that they can reach areas of cerebral interest also in physiological conditions.

Concerning their production, ACN is obtained as a by-product from the synthesis of acrylonitrile (SOHIO process, catalytic ammonium oxidation of propylene), which is performed in gas-phase with metal oxide catalysts using ethylene and/or propylene, ammonia and oxygen [59]. The PC is mainly prepared by propylene oxide ring-opening in a CO2 atmosphere under harsh conditions of pressure and temperature [60,61]. Therefore, both most used solvents to prepare electrolyte are synthesized from non-renewable feedstock.

In this work we have synthesized a panel of substances (LA-H,H; LA-H,Me; LA-Me,Me) obtainable from renewable sources and with a potentially benign ecotoxicological profile, and we used them for the first time as solvent for non-aqueous EDLCs.

Based on preliminary results, the solvent LA-H,H displayed the highest conductivity (8.5 mScm−1) and was therefore chosen as the best candidate for subsequent characterizations. Two electrolyte-based LA-H,Hs and PCs were prepared with the same conducting salt (TEMABF4) to investigate any differences related to solvent effect. The prepared electrolytes were used to assemble symmetrical EDLC, which were thoroughly characterized by evaluating:


The ensemble of results provided by the assembled devices highlights the relevance of the solvent-salt interactions to determine the overall performances. Compared to PC based electrolytes, at 2.6 V LA-H,H/TEMABF4 showed approximately the same coulombic efficiency in the range of 94%–95% and a modest increase of capacitance retention with a high scan rate. The overall storage performances achieved with LA-H,H solvent were adequate for EDLC application (Esp > 10 Whkg−<sup>1</sup> and Psp > 20 kWkg<sup>−</sup>1), and comparable with those achieved with PC/TEMABF4. EIS analysis confirmed small variations among each resistance parameters, and a more prominent difference between the time. Interpreting this behavior with the storage parameters recorded at a high specific current, it is possible to assume that the decrease in specific energy and specific capacity for LA-H,H/TEMABF4 is due to an incomplete formation of the electric double layer, since at high current density the charge time was below the relative time constant. Final stability investigations performed with LA-H,H/TEMABF4 revealed high performance, as capacitance retention (never less than 90% even at high current density) and coulombic efficiency were close to 100% for all the cycles. However, this new electrolyte should be cycled over a longer cycling period (at least 10,000 cycle) to definitively validate its stability in EDLC devices and will be the aim of future developments.

In this work, it was therefore demonstrated that compounds based on the 5-methyl-1,3-dioxolan-4-one scaffold (Figure 6) are sustainable non-aqueous solvents for applications in energy storage devices. In the future, this class of electrolytes will be extended to other modulable and still unexplored solvents by using other common α-hydroxy acids, such as glycolic, mandelic and 2-hydroxyisobutyric acid, as starting material.

**Figure 6.** Generic structure of 2,2-R,R -5,5-X,X -1,3-dioxolan-4-one compounds.

#### **5. Conclusions**

Most of the literature articles on non-aqueous electrolytes aims to uniquely increase the EDLC's performance, often neglecting issues of great importance such as sustainability and safety, and relegating the relevance of these issues only to aqueous electrolytes. In this work we have subverted this concept designing aprotic polar solvents from renewable sources, such as green lactic acid, and investigating their performances.

Within this framework, three dioxolanes with small structural variations in position 2 were synthesized: 2,2,5-trimethyl-1,3-dioxolan-4-one (LA-Me,Me), 2,5-dimethyl-1,3 dioxolan-4-one (LA-H-Me), and 5-methyl-1,3-dioxolan-4-one (LA-H,H). As expected, small structural variations significantly influence their performances, such as conductivity and solvent ability vs TEMABF4, which allowed us to select 5-methyl-1,3-dixolan-4-one for a thorough characterization of the EDLC device. The results demonstrate that the LA-H,Hbased electrolyte is suitable for the application, and competitive with the one based on commercial PC. In fact, an operational potential typical of non-aqueous electrolytes (2.6 V, CE ≈ 95%) and adequate storage parameters have been reached (Esp > 10 Whkg−<sup>1</sup> and Psp > 20 kWkg<sup>−</sup>1). This result paves the way for the use of a wide class of solvents based on α-hydroxyacid ketals as sustainable alternatives to those obtained from non-renewable fossil sources. Moreover, thanks to their structural versatility, the room for improvement is still large and further studies will be aimed at increasing performances by finely tuning the solvent-salt combination.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/en14144250/s1, Figure S1: 5-methyl-1,3-dioxolan-4-one (LA-H,H) characterization, Figure S2: 5,2-dimethyl-1,3-dioxolan-4-one (LA-H,Me) characterization, Figure S3: 5,2,3-trimethyl-1,3-dioxolan-4-one (LA-Me,Me) characterization, Figure S4: Triethylmethylammonium tetrafluoroborate (TEMABF4) characterization, Figure S5: ESW investigation, Figure S6: Operative voltage investigation, Figure S7: Nyquist plot analysis and resistances evaluations, Table S1: ESW data at different current densities cut-off, Table S2: Relevant performance from cited literature.

**Author Contributions:** Conceptualization, M.M. and F.R.; formal analysis, M.M.; investigation, M.M.; data curation, R.E.; funding acquisition, M.D.S., G.A.; supervision, A.L., A.B., F.R.; writing-original draft preparation, M.M., R.E., M.D.S., A.L., A.B., F.R.; writing-review and editing, M.M., R.E., M.D.S., G.A., A.L., A.B., F.R. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** Pier Paolo Prosini (ENEA DTE-SPCT, C.R. Casaccia, Santa Maria di Galeria, Rome, Italy) and his research group are acknowledged for their valuable contribution and support on coin cell preparation.

**Conflicts of Interest:** The authors declare no conflict of interest.
