Prototype Aqueous Lithium-Ion Battery with a Lithium-Ion Conductive Solid Electrolyte Separator

Abstract

Lithium-ion batteries (LIBs) with a non-aqueous electrolyte are capable of high energy density, high power output, and long cycle performance. These batteries are regarded as the most promising sources of power for electric vehicles (EVs), hybrid electric vehicles (HEVs), and grid storage. However, current LIBs pose a safety risk due to their flammable organic electrolytes. In contrast, lithium-ion batteries with aqueous electrolytes are safer and cost less than current LIBs, but the electrochemical stability window of aqueous electrolytes is narrower than that of the organic electrolytes used in LIBs. The stable operating voltage window of an aqueous electrolyte is approximately 1.23 V thermodynamically, beyond which the electrolysis of H2O occurs along with the evolution of H2 or O2 gas. As illustrated by a Pourbaix diagram, the overall electrochemical stability window of aqueous electrolytes is constant and independent of the pH of the electrolytes. In this study, we report the widening of the overall electrochemical stability window with two electrolytes having different pH values. We utilized a lithium-ion conductive solid electrolyte to separate the electrolytes into the cathode and anode sides of an aqueous lithium-ion battery in order to suppress the evolution of oxygen and hydrogen. In particular, we fabricated an aqueous lithium-ion cell with a lithium-ion conductive solid electrolyte as the separator between two inorganic lithium salt layers. A spinel lithium manganese oxide (LMO) was used as the cathode active material while a spinel lithium titanate oxide (LTO) with a Li+ insertion potential of 1.55 V vs. Li was used as the anode active material, leading to less hydrogen evolution compared with the graphite anodes used in current LIBs. A lithium-ion ceramic plate (LICGC™, Ohara Inc.) with a thickness 150 µm and conductivity 10-4 S/cm at 25°C was used as the separator. The plate was able to separate the different electrolytes into the cathode and anode sides. Then, 2 mol/L Li2SO4 and 12 mol/L LiCl, both of which have higher conductivity than the organic electrolytes of current LIBs, were used as the cathode and anode electrolytes, respectively. The cell was constructed with an LMO cathode electrode (10 mm in diameter, 0.1 mm thick), an LTO anode electrode (approximately the same dimensions as the cathode), the solid electrolyte separator (2.5 ´ 2.5 cm, 150 µm thick), and an aqueous solution with 2 mol/L Li2SO4 and 12 mol/L LiCl as the cathode and anode electrolyte, respectively. The charge/discharge current density was 0.282 mA/cm2 at the 1C rate. Figure 1 shows the charge and discharge curves of the LMO/LTO cell at the 20th cycle (25°C). The aqueous test cell exhibited an average discharge voltage of over 2.4 V and a coulombic efficiency of over 94%. Figure 2 shows the cycling stability and coulombic efficiencies of aqueous LMO/LTO cells at the 1C rate. To prevent drying of the electrolyte in a long-term experiment, 12 mol/L LiCl was used as electrolytes for both the cathode and anode. The retention of discharge capacity and coulombic efficiency was approximately 100% after 1000 cycles. The coulombic efficiency of the initial cycle, in which the pH of the 12 mol/L LiCl electrolyte was neutral (measured with pH test paper), was relatively low. During the 20th cycle, which had relatively low coulombic efficiency, the anode electrolyte changed from acidic to basic and hydrogen evolution with electrolysis of water was suppressed. After the 20th cycle, the electrolyte pH at the anode electrode surface was close to 14, whereas the entire electrolyte pH was close to neutral. The pH of the local electrolyte, situated between the anode electrode surface and the solid electrolyte separator, changed from acidic to basic, resulting in the suppression of oxygen and hydrogen gas generation without changing the pH of the cathode electrolyte. The initial coulombic efficiency was dependent on the anode electrolyte pH due to the lower coulombic efficiency of the anode compared with the cathode. Additional electrochemical tests will be reported in the future. Figure 1