0x62448568…9229sent to0x8011df58…516c·#24,270,008·view on Etherscan
Water adsorbs to surfaces it can hydrogen bond to, and in the adsorbed phase the mobility of the ions that form by auto-ionization is asymmetric, the hydroxide ion is stuck within the adsorbate while the hydrogen ion is free to move. Thus, the hydrogen ion spreads out by diffusion, in equilibrium with the drift current from the resulting electric field (backwards towards the then negatively charged adsorbate). If there is dioxygen (O2) in the surrounding water, hydroxide ions occasionally break down into electrons, dioxygen and water, and the electrons move to combine with the hydrogen ions and dioxygen to form water. This consumes one water and produces one water - thus there is no change in quantity within the container. But this same process can happen over a membrane, into another container. If the surface that water is adsorbed to is permeable to protons and electrons, and there is dioxygen on the other side of the surface, protons will tend to diffuse over the membrane and the dioxygen and protons will occasionally accept electrons from hydroxide ions that break down in the adsorbate on the other side, thus one water is produced in the new compartment and one water is lost in the previous compartment. The same reaction that spontaneously happens within the adsorbate and bulk water in a compartment - when you add a separate compartment, will tend to move water. Naturally, if you have adsorbate on both sides of the membrane, the side with more adsorbate will tend to more frequently spontaneously produce water on the side with less adsorbate, compared to the side with less adsorbate. Adsorbate, like ice, is impaired by salt, so if you have salt water in one compartment and fresh water in the other, you will tend to produce water in the saltwater compartment and consume it in the freshwater compartment.