Kyle Knust at al. explain here how they intend to desalinate sea water not with established methods such as multi-stage flash distillation or reverse osmosis but by electrochemistry.
In their proof of concept the anodic pole of a bipolar electrode sits in the middle of a junction of a branched microfluidic channel. Sea water spiked with fluorescent tracer tris(bipyridine)ruthenium(II) chloride is fed into the system and voltage (3 V) applied. Interestingly the channel partially blocked by the anode is 25% less saline (by conductivity) and the fluorescent tracer material also opts for this channel.
The rationale? At the anode chlorine ions are oxidized to chlorine gas creating a local "ion depletion zone" and a electric field gradient motivating ionic species to move away from it.
Entirely happy? What is the "ion depletion zone" exactly. As chlorine ions get oxidized, sodium ions accumulate causing all incoming ions to reroute. You might as well call it an accumulation zone. Another matter: the bipolar electrode was at one time a distinct research interest of Martin Fleischmann, you know the guy we are supposed ridicule for 20 years to come and then revere as a genius. This Daniell cell without a bridge also has the look and feel of electrolysed water. More explanations are in order but the paper does not provide them. Finally, being a microfluidic device the output is only 0.04 microliter per minute. So there is some way to go.
But at least we have the movie!