Here is a link to a 3 year old essay of mine about how membrane reactors work and how you can use them to get past the normal limits of thermodynamics. The words are good, as is the example application, but I think I can write a shorter version now. Also, sorry to say, when I wrote the essay I was just beginning to make membrane reactors; my designs have gotten simpler since.
Above, for example, is a more modern, high pressure membrane reactor design: 72 tube reactor assembly; high pressure. The area at right is used for heat transfer. Normally the reactor would sit with this end up, and the tube area filled or half-filled with catalyst, e.g. for the water gas shift reaction, CO + H2O –> CO2 + H2. According to normal thermodynamics, the extent of reaction for this will not be affected by pressure once it reaches equilibrium, only by temperature. If you want the reaction to go reasonably to completion, you have to operate at low temperatures, 250- 300 °C, and you have to cool externally to remove the heat of reaction. In a membrane reactor, you can operate at much higher temperatures and you don’t have to work so hard to remove heat. The trick is to operate with the reacting gas at high pressures, and to extract hydrogen at lower pressures. With a large enough difference between the reacting pressure and the extract pressure, you can achieve high extents (high conversions) at any temperature.