I want to calculate the surface tension of beads. But after my model was simulated by DPD, the vacuum layer disappeared. 

The atomic model I built in a similar way（cleave surface，build vacuum slab） retains the vacuum layer after the NVT equilibrium.