Among energy storage technologies beyond lithium-ion, alkali metal-oxygen batteries stand out for their unparalleled theoretical energy densities. Based on the reaction between dioxygen and alkali metal ions in non-aqueous electrolytes, metal-oxygen batteries could store up to 10 times more energy per unit mass (in the case of Li-O2) when compared to today’s lithium-ion batteries1. Despite extensive efforts elucidating the mechanism of the reaction between alkali cations and dioxygen, this technology is still far from maturing into a commercial device. The main drawbacks of Li-O2 systems are electrolyte and cathode instability, low conductivity of lithium peroxide (Li2O2) leading to poor cycleability. In contrast to Li-O2, potassium-O2 discharge reactions proceeds via a one electron reduction leading to potassium superoxide (KO2) as the sole discharge product2-5. Here, different carbon based cathodes were investigated in K-O2 cells. Furthermore the stability of the K metal anode with different non-aqueous electrolytes was compared. The cell capacity varied depending upon the cathodic carbon structure employed (Figure 1a). KO2 was the only discharge product identified via Raman spectroscopy (Figure 1b) and powder X-ray diffraction. FTIR spectra taken of the cathodes after first discharge, showed no evidence of carbonate species or other parasitic reaction products. These results are encouraging, since the large overpotential observed within Li-O2 cells was found to be related to side reactions on the carbon electrode surface6. The voltage gap in K-O2 cells was lower than 100 mV, and round trip efficiency, as high as 98%. The cycleability of the cells was found to be limited by side reactions between the potassium metal anode and ether-based electrolyte in the presence of dioxygen reduced species, rather than side reactions within the carbon cathode. The main components of this insulating anode surface layer were K2CO3, KOH and KO2. Similar anode surface layer composition was found for different combinations of ether electrolytes and potassium salts. However, while using super-concentrated electrolytes (3M and 5 M KTFSI in dimethoxyethane), O2 crossover was inhibited, and no KO2 was detected on the potassium metal anode surface. As a result, these cells were able to be cycled over 50 times under shallow cycling of 50 mAh/g. The superconcentrated electrolytes had no effect on the cathode reaction product, and the battery cycled through KO2 formation and oxidation. Capacity fade eventually occurred, as repetitive stripping/plating of K metal anode caused the disruption of the protective layer. This work highlights that once a suitable electrolyte has been selected, a metal-O2 cell can indeed be cycled numerous times. The challenge is to increase the cycleable capacity to more appealing values, whilst maintaining cell stability. [1] P.G. Bruce, S.A. Freunberger, L.J. Hardwick, J-M. Tarascon, Nature Mater. 11, 19 (2012). [2] X. Ren, Y. Wu, J. Am. Chem. Soc. 135, 2923 (2013). [3] X. Ren et al. ACS Appl. Mater. Interfaces 6, 19299 (2014). [4] W.D. McCulloch et al. ACS Appl. Mater. Interfaces 7, 26158 (2015). [5] X. Ren et. al. Adv. Energy. Mater. 1601080 (2016). [6] B.D. McCloskey et. al. J. Phys. Chem. Lett. 4, 2989 (2013). Figure 1