Lithium–oxygen batteries are promising but have many challenges. Unlike lithium-ion batteries, they are usually discharge–charge cycled with capacity cutoff instead of potential cutoff, which brings controversy. Additionally, which superoxide intermediate, the dissolved or the adsorbed superoxide, is more reactive and leads to cell premature death and unsatisfactory discharge capacity? These questions puzzle researchers and impede the development of lithium–oxygen batteries. Herein, on one hand, we tried to decouple the influence of discharging potential and discharging current density on the discharge products and side reactions. We found that the electrode potential has more impact on the side reactions than the current density. The low potential leads to a high ratio of Li₂CO₃ to Li₂O₂ in the discharge product and hence more surface passivation. On the other hand, to identify the more reactive and aggressive species that cause surface passivation, a flow cell setup was applied to suppress the solution route and maximize the products from the surface route. Results show that more Li₂CO₃ was identified under a large flow rate and thus the intermediates in surface route appear to be more reactive than that in solution route.