You are not the first person to think of this... Peter Watts uses thid kind of oxygen source in his underwater fiction, including _Starfish_ ([full story available on his website for free][1]) and in a short story [snippet][2]. The rifters in _Starfish_ had a lung replaced by a system that does the appropriate gas exchange itself, removing all the problems of bad thing dissolving into and out of bodily fluids under extreme pressure, and the latter has a simpler system that produces breathing gas that is inhaled by an unmodified human. I can't tell which system you're specifically interested in, though note that the former is more invasive and needs a direct gas exchange system with blood pumped through it (a little like [ECMO][3]) to get oxygen in and CO<sub>2</sub> out, whereas the latter uses the lungs for this purpose (effectively [hydrox][4] diving). That's obviously simpler and uninvasive, but doesn't solve problems of [gas narcosis][5] or [the bends][6]. I won't go into further detail here, but obviously there are details you need to care about. Lets consider your design ideas though: > I've elected to place them in large "wings" on their backs A professional aerobic athlete might be breathing at 150 litres of air per minute at sea level whilst working hard, and this seems like an OK "maximum output" level. About 5% of each lungful is consumed, giving you a peak O2 flow of only an eighth of a litre per second. The [density of oxygen at STP][7] is about 0.04013 mol/dm<sup>3</sup> (a little less than the equivalent ideal gas)), so we can see that our athlete needs ~5mmol of oxygen per second. There's a single oxygen atom per water molecule. We therefore need 10mmol of water to general 5mmol of O2. The molecular weight of water is 18.01528 g/mol, therefore you can get your entire oxygen needs from .18g of pure water per second... an equivalent flow of just a 10ml per minute! (and FWIW, my results tally with [this chemistry.SE][8] question, so my working can be cross checked!) Obviously flow rates in a real devise would likely be higher, because you don't want your electrolysis cell to go completely dry to avoid clogging from dissolved materials like salt or calcium carbonates. You can see that you can supply all the water you need for your electrolysing-gill needs from a tiny pump, with a tiny inlet hole (or several holes to reduce the risk of blockage). A resting human needs more like 5-8l/m of air, which is vastly less than the athlete, and so someone floating or paddling gently will need an order of magnitude less water. **There's no need for huge surface areas for extracting water by electrolysis**. > And apparently saltwater is a good conductor? It is, but we've established that your electrolyser can be pretty small and hence well protected. If necessary it could be run in a pulsed mode of operation where the water being electrolysed is electrically insulated from the outside by valves, and when waste hydrogen is burped out, water can be sucked in to refill the chamber. Electroceptive species like sharks might notice your device running, but whether they'd find it interesting or unpleasant I couldn't say. Power requirements are a slightly larger problem. Water needs [237.24 kJ/mol][9] to split, so to split the required rate of 10mmol water per second **you need a ~2380W power supply**, which at the necessary [1.23v][10] equates to a nearly 2KA current. This is equivalent to the power supply needed for welding. You'll be wanting a decent fuel cell or battery pack to drive your air supply! Of course, a _normal_ person might need as little as a tenth or a twentieth of the pro-athlete's flow rate, and in turn this reduces power demands to a more achievable **238W**, leaving you with power demands more like an electric bicycle which is obviously acheivable with modern day tech without being too bulky. Note: if you are generating gas to be inhaled you will need to split more water to provide enough hydrogen to dilute the oxygen, because breathing to high a partial pressure of oxygen results in [acute oxygen toxicity][11] which can be fatal. This means you'll generate an excess of oxygen that will have to be vented into the sea. In any case, future fuel cells should give you pretty good energy density, and future advances in artificial photosynthesis should provide the fuel which can be conveniently made at sea on the surface. Some further chemical cleverness will doubtless be required. For example, electrolysis of salt water can release chlorine gas which you probably don't want to breathe, and chloride ions can damage the electrolysis equipment. Some research [already exists][12] in this area, though it is worth considering that some kinds of damage (or sabotage!) to your electolytic lungs in seawater could still cause chlorine to be released. [1]: https://www.rifters.com/real/STARFISH.htm [2]: https://www.rifters.com/crawl/?p=7138 [3]: https://en.wikipedia.org/wiki/Extracorporeal_membrane_oxygenation [4]: https://en.wikipedia.org/wiki/Hydrox_(breathing_gas) [5]: https://en.wikipedia.org/wiki/Hydrogen_narcosis [6]: https://en.wikipedia.org/wiki/Decompression_sickness [7]: https://www.engineeringtoolbox.com/oxygen-O2-density-specific-weight-temperature-pressure-d_2082.html [8]: https://chemistry.stackexchange.com/a/15879/105853 [9]: https://en.wikipedia.org/wiki/Electrolysis_of_water#Thermodynamics [10]: https://en.wikipedia.org/wiki/Electrolysis_of_water [11]: https://en.wikipedia.org/wiki/Oxygen_toxicity [12]: https://www.pnas.org/content/116/14/6624