Some difficulties would be present to obtain the velocity necessary to overcome the recoil. My proposal would be to let the gas of atoms escape through a nozzle. Ss the volumetric flow rate has to be the same before and after the nozzle, and the nozzle has a very small diameter the velocity of the atoms after the nozzle would be much higher if the diameter stays the same.