State of the art electrolytes for lithium ion batteries (LIBs) consist of a lithium salt dissolved in a mixture of cyclic and linear carbonates, mainly using ethylene carbonate (EC), dimethyl carbonate (DMC) and/or diethyl carbonate (DEC). Notwithstanding many beneficial properties, e.g. high conductivity and low viscosity, the thermal operation window is limited especially in regards to low volatility and flammability. To overcome well-known drawbacks of carbonate based solvents, alternative solvents are desired to simultaneously expand the operating temperature range, improve the safety and maintain excellent electrochemical performance. As conductive salt, LiPF₆ is still the lithium salt of choice. However, it suffers from drawbacks related to trace amounts of water as well as its decomposition at high temperatures. Lithium bistrifluoromethanesulfonimide (LiTFSI) represents a promising alternative salt for LIB technology to possibly overcome the aforementioned challenges. However, an additional challenge in LITFSI based electrolytes is set on the dissolution of the aluminum current collector at already moderate potential values (<4V vs. Li/Li⁺) when using carbonate based solvents.

The goal of our research was to develop new safer electrolyte formulation based on non-carbonate based single solvents with LiPF₆ or LiTFSI as conducting salt. In this work, an ester functional group was introduced to the nitrile molecular structure to support salt dissociation and to enhance the compatibility towards lithium metal and graphite based anodes. The basic framework for the possible structures was either R1-COO-R2, whereas two different linkages of the nitrile moiety were possible. In the first one, nitrile functionality is positioned on the acid part of the ester, whereas in the second it is situated on the alcohol part. In order to select the best possible combination of ester and nitrile moieties, a computational screening by means of theoretical calculations of the main physical and electrochemical properties was preceded. Over 2000 possible structures were calculated in regards to important parameters namely viscosity and pkA value as well as the melting point. All possible molecules were divided into the two possible linkages. Methyl 3-cyanopropanoate (MCP) and, methyl 2-cyano-2-methylpropionate (MCMP) were identified as the two most promising candidates for the acid linkage. Two additional compounds, namely methyl cyanoacetate (MCA) and ethyl-2-cyano-2-methylpropionate (EMCP) were included for better evaluation of the results obtained by theoretical calculations. For the second possibility, three molecules were synthesized, namely 2-cyanoethyl acetate (CEA), 2-cyanoethyl 2,2,2-trifluoracetate (CFA) and 2-cyanopropan-2-yl acetate (CYA) as depicted in Figure 1.

In the frame of this work, the newly proposed electrolyte solvents were analyzed in regards to their liquid window, viscosity and conductivity. In our study, a very good agreement between theoretical and experimental results was obtained. Additionally, all possible solvents were screened with respect to their compatibility to graphite based anodes. We found that it is necessary to use a film-forming additive. This study showed that possible acid linked candidate proved to be MCP and the best alcohol linked proved to be CEA, which were not the ideal candidates in regards to their physicochemical properties. After having identified the two most suitable candidates, experiments in full cell setups with 4 V cell chemistry were performed. As we used a single solvent approach, the obtained results were compared with 1 M LiPF₆ in PC based electrolytes containing FEC as film forming additive and additionally to a state of the art electrolyte: 1 M LiPF₆ in EC:DMC (1:1 by wt.). These experiments showed that at moderate cycling rates of 1 C the proposed electrolytes perform comparable to the carbonate based electrolytes. To further evaluate the thermal properties of the electrolytes, Heat-wait-search (HWS) tests were performed to determine the onset temperature value for a thermal runaway. Furthermore, the compatibility of MCP and CEA with the alternative conductive salt LiTFSI was investigated, showing that the cyanoester based solvents show negligible to no corrosion behavior on the aluminum current collector in comparison to the carbonate based solvents.

In conclusion, we present a complete study obtained from selected theoretical calculations of a promising new class of solvents for the implementation of MCP and CEA in full cell set-up as possible alternative solvents for LiPF₆ and LiTFSI. Comparable thermal and electrochemical behavior in comparison to a one solvent carbonate based system and better thermal behavior in comparison to a two solvent carbonate based system was recognized.