▎ 摘　　要

Graphene oxide (GO) membranes have been demonstrated to possess promising prospects in pressure-driven desalination, in which the desalting performance is generally investigated in a dead-end flow system by following the test procedure for polymers. However, there exists a blind zone in desalination test whether the asreported "retention" is reliable, as there is no experimental evidence to support that the stable salt rejection has been achieved. On this basis, for the first time, we investigated the long-term desalting performance of GO membranes on five different salts, including Na2SO4, NaCl, KCl, MgSO4, and MgCl2, in a conventional dead-end flow device by following the two typical experiment processes for GO membranes. Surprisingly, it was found that the salt rejection of GO membranes dropped down to 0 with the prolongation of testing time. The result indicated that GO membranes were not capable of rejecting salts steadily under pressure-driven conditions. The "salt rejection" in present reports, is only the initial highest value based on their test method. In this study, the underlying mechanism was also elucidated. With extending test time, the metal cations in salt solution would be gradually adsorbed on GO membranes, which shields the negative surface potential, thus leading to the loss of desalination performance. This finding not only rationalizes the challenge of GO membranes in pressure-driven desalination in a dead-end flow system but also prompts us to reassess the feasibility of GO membranes in desalination applications based on Dorman exclusion.