Temperature-Dependent Gassing Analysis By on-Line Electrochemical Mass Spectrometry of Lithium-Ion Battery Cells with Commercial Electrolytes

Abstract

Li-ion batteries (LiBs) are known to evolve gases during their first few formation cycles due to the formation of a solid electrolyte interphase (SEI) at the anode active material particle surface. Simultaneously, the electrolyte can react by other mechanisms leading to the formation of volatile species. 1,2 At a high state-of-charge, the cathode can also release reactive oxygen which causes the formation of CO and CO2 from the oxidative decomposition of the electrolyte.3,4 Consequently, by the analysis of evolved gas species one can study the effect of electrolyte additives, reveal substantial degradation processes within LiBs, and rationalize optimized formation strategies.4–6 The evolution of gaseous species is typically quantified in-situ by monitoring the cell pressure (e.g., via the so-called Archimedes principle) or by online mass spectrometry.5,7 Concerning the latter, an on-line electrochemical mass spectrometer (OEMS) was developed for the application in LiB research and proved to be particularly powerful, since it allows a quantitative analysis and the identification of the evolved and consumed gas during the operation of a battery.4,7 The quantification of OEMS data, however, is challenging when the vapor pressure of the electrolyte is high, which, e.g., is the case for linear carbonates, particularly at elevated temperatures. Mixtures of linear and cyclic carbonates are widely used in LiB applications due to their low viscosity, high salt solubility, and high conductivity. In contrast to the cyclic ethylene carbonate (EC) with a vapor pressure of 0.25 mbar, the vapor pressure for the typically used linear carbonates dimethyl carbonate (DMC), diethyl carbonate (DEC), and ethyl methyl carbonate (EMC) can vary from 23 to 80 mbar at room temperature.8 Their vapor pressure further increases by a factor of ~3 when increasing the temperature from room temperature to 45 °C.9 Thus, a significant share of the gas in the head-space of the OEMS cell consists of electrolyte vapor. As they undergo fragmentation in the mass spectrometer, a background signal is superimposed on the relevant mass traces related to hydrogen, ethylene, carbon monoxide, oxygen, and carbon dioxide.1 Furthermore, the cell pressure inherently decreases due to the continuous leak through the capillary (with a leak rate of ~1 µL/min) that connects the OEMS cell (~10 mL volume) with the mass spectrometer. Because of the liquid reservoir of electrolyte in the OEMS cell and its constant vapor pressure, this results in an enrichment of the gas phase with electrolyte vapor. Analyzing the pressure-dependent flow of gaseous species into the mass spectrometer allowed us to introduce corrections for signal normalization, electrolyte background change, and signal quantification.In the present study, we applied these corrections in the investigation of the influence of LP57 (1 M LiPF6 in EC:EMC 3:7wt/wt) and LP47 (1 M LiPF6 in EC:DEC 3:7wt/wt) electrolyte on the gassing characteristics during the formation of full-cells with a graphite anode and a Ni-rich cathode from low (10 °C) to high (45 °C) temperatures. From that, we were able to conclude how the formation temperature influences the gas evolution related to SEI formation as well as the trans-esterification of the electrolyte and the lattice oxygen reactivity towards different electrolytes. A comparison to a model electrolyte (1.5 M LiPF6 in EC) with a low vapor pressure (< 0.25 mbar) allowed us to assess the detection limits of all the signals depending on the electrolyte’s vapor pressure and to evaluate the validity of our proposed signal corrections. Acknowledgment: The authors thank BMW AG for their financial support.