Aren't they all sunny at the top of the atmosphere?
> gives about 1360 watts per square meter
I assume that's very much latitude dependent
There’s also more atmosphere in the way the further you go from the equator which will affect how much solar radiation reaches the earth. Overall only about half of solar radiation reaches the surface.
What would the math look like if they aren't necessarily going to boil?
Of course on the flip side, your goal is fresh liquid water, so you need to condense the vapor. Condensing hot vapor is easy, just expose it to cooler ambient conditions. Condensing ambient vapor is harder, and will require you to run something like a refrigeration cycle or a chemical desiccant system which will need energy to be regenerated.
Most commercial systems use vacuum distillation which boils water at low temperatures and pressures, which has its own drawbacks but is generally more efficient.
You are incorrect that at ambient temperature the lower energy bound is set by the latent heat of vaporisation, as others have pointed out this is theoretically recoverable.
At ambient temperature there is however still a fundamental physical limit: the solvation energy of the salt in the water: https://en.wikipedia.org/wiki/Solvation#Solvation_energy_and...
Now the whole globe does not have the same ambient temperature, and as you know about global warming it would be great to shed some energy in the form of heat.
There are many forms of desalination. Another way to desalinate is freezing: when salty water freezes, it pushes out the salt, so while desalinated water ice forms, the liquid water surrounding the ice will increase in salinity and become brine. One could then use simple nets or grills to separate ice from brine.
Suppose one has a space elevator, or even a tether from a balloon, but capable of carrying significant weight.
The temperature falls roughly adiabatically with height. Above the tropopause the atmosphere is essentially cloud-free, CO2 free and below freezing point (say -60 deg C). Hence the latent heat of fusion (freezing) can be shed to outer space. So it should be possible to lift salty water up an elevator, allow it to freeze over, separate brine from ice at the top, then lower the separated brine and ice.
The energy required to lift the brackish water is compensated by the energy released by lowering the freeze-distilled water and brine. What comes up must go down, so simplistically speaking a pulley in equilibrium, so that the only energy intentionally exerted is lost to pulley and air friction. Then one would be cooling the planet and receiving frozen freeze-distilled water at the same time.
The law of conservation of misery is typically not a fundamental law of nature, but imposed by reluctance to study of those who dictate artificial laws.
With regard to the energy expenditure for reaching that temperature: if we were merely raising the temperature of the water and then cooling it down again, I think it would be correct to say that with a completely efficient contra-flow heat exchanger, perfect insulation, and no pumping losses, the steady-state heat input could be arbitrarily low.
If we now modify this system to evaporate and then condense some of the water at the point of highest temperature, we would have to supply, and then extract, the latent heat at whatever temperature the evaporation is performed at. Once that has been performed, the outflow would comprise of the same amount of water as before (and at the same temperature), and it would be equally available to warm up the incoming stream as in the initial scenario (though now we would need two heat exchangers in order to keep the fresh water separate.)
Of course, both the heat exchanger and the insulation will have losses, but we are rejecting quite a bit of heat in the condenser, and it is at the highest temperature in the system. Would that, in principle, be available to make up for any losses elsewhere? This makes me wonder if, counter-intuitively, it could be more efficient to do the distillation at higher temperatures, at least up to the point where the diminishing latent heat can no longer compensate for the losses of running at a higher temperature?
I'm leaving out some considerations that I don't know how to handle (and probably others that have not occurred to me.) For one thing, there's the question of what happens if the evaporation occurs into a chamber containing some air, rather than just steam (my guess is that the relevant temperature is determined by the water vapor partial pressure.) For another, what difference does having salt dissolved in the water make? And this may all be moot, as this system has no moving parts, so the pressure is probably atmospheric (or somewhat below, if the condensation can be exploited to create a partial vacuum.)
I have no idea if any of this makes the slightest bit of sense, and it's probably wrong - as you say, most systems run at reduced pressure.
[1] https://www.engineeringtoolbox.com/water-properties-d_1573.h...
With perfect insulation and heat recovery (zero loss) all that matters is the change in entropy between the starting and end products. Both the energy for raising the temperature and for the vaporization is theoretically recoverable (when you condense a vapor back to a liquid it releases the same amount of heat that it took to vaporize it). But you can't have perfect insulation and heat recovery in practice, and the losses become worse with increasing temperature - or more accurately increasing temperature difference, so trying to cool things down below ambient won't help you either.
I Don't think that can be right - for example, at 18 °C, the isobaric specific heat of water is 4.18 kJ/(kg.K), while the rate of change of the latent heat with temperature is only 2.4 kJ/(kg.K), and at 100 °C, the figures are 4.22 kJ/(kg.K) and 2.7 kJ/(kg.K) respectively.
> Both the energy for raising the temperature and for the vaporization is theoretically recoverable (when you condense a vapor back to a liquid it releases the same amount of heat that it took to vaporize it).
Only up to a point: you cannot condense steam at 100 °C in a condenser where the incoming coolant is also at 100 °C. Using only passive methods (heat exchangers) you cannot, even assuming perfect efficiency, recover all of the heat needed in a distillation process for reuse within that distillation process.
While distilling at higher temperatures need not be anywhere near as inefficient as it seems if you don't include the use of heat exchangers, the numbers given above don't seem different enough to justify distilling at a higher temperature than necessary, under realistic assumptions of efficiency, which is not surprising, given that it does not seem to be done.
You use the cold incoming water to condense the hot vapor, thus pre-heating it prior to distillation. You get out all the heat you put in, the issue is that vapor has higher entropy than liquid water so you can't use that heat efficiently enough to vaporize the same quantity of water. That's in the efficiency term we are handwaving away.
If it works less well on cool days, well you don't need as much water on cool days as you do on hot ones.
[1] https://www.sciencedirect.com/science/article/abs/pii/S13594...
Edit: If you're thinking of something like an ultrasonic humidifier, I don't think these actually evaporate the water[2]. The mist these produce would still contain salt if you tried to use them for desalination.
[2] https://en.wikipedia.org/wiki/Humidifier#Ultrasonic_humidifi...
Aside:
I have a home-built ultrasonic humidifier. If I run it with Boulder, CO tap water that is low in TDS, it only takes a day or so to have a PM2.5 >600 in my house. For this to work I had to install an RO filter in order to humidify with ultrasonic and not degrade air quality.
I didn't know that legionella can kill macrophages from the inside. That's nasty.
There are definitely efficiencies to be had, though I don't know enough of the math to judge one vs. the other. During my brief patent career, I wrote the patents for a distillation system where the main elements involved heating water that was distributed across rotating blades (heat + surface area + air movement) to evaporate the water. When the water was collected, it passed through a heat exchanger that exchanged heat with the in-flowing water. The result was a very efficient system on a small scale, at least.
*edit: although window and wall unit HVAC use the blades to fling water around so the condenser gets the coolest possible air
Essentially, they find an equilibrium between the cold water coming in, warming in the sun, an increasing amount evaporating into the warm, damp chamber, and the remaining brackish water being cooled by the new water.
> if each stage were scaled up to a square meter, it would produce up to 5 liters of drinking water per hour
so 1m², times the number of layers.
If heating water is the goal then going directly to heating the water is more than 4x the power capture of a solar panel. Dropping one solar panel on your roof for a water heater pays off its just that the solar heaters are kind of expensive dealing with all the other aspects, like if the pump fails or its too cold and the water would freeze etc etc. Those other aspects are probably going to dominate the complexity and cost of the devices.
(Not affiliated, no idea if that particular one is any good, it was just one of the first results)
A solar thermal heater typically lasts 10 years, its a pretty harsh environment and they have pumps and expansion vessels and water in the heat range of -25 through +80 C is just damaging over time. Where as a solar panel has a rated 80% output at 25 years. That mismatch is problematic for Solar panel payoff, 10 years is only just past break even point. Whereas the solar thermal heaters typically pay off within a few years so the reduced lifetime isn't such a big problem (but both are driven by local gas and electrical prices).
Companies are trying to do this better and products exist to do it. The added installation complexity and reduced expected lifetime are all a bit of an issue. Currently I suspect most people are better off with a dedicated solar heater and solar panels separately, at least for now until the water channels in the panels at least are reliable enough that they meet the panel lifetime and then the pump is installed in the roof cavity so its cheap and easy to replace. Work still needs to be done to work out the details and make the parts that fail easily swappable.
Right now most people are better off (from a total cost of ownership perspective) with PV panels + heat pump water heater. The latter can achieve a COP of 3.0 (300% "efficiency"), and also can just use grid electricity on a cloudy day. Unless you are off-grid, the additional cost/complexity of running plumbing to the roof for a solar water heater doesn't pencil out.
Some premium hot water heaters like the SanCO2 can achieve COP of up to 5.0 [2], so high that they don't need backup resistance elements, even in northern climates.
1. https://www.energystar.gov/productfinder/product/certified-w...
2. https://www.smallplanetsupply.com/small-planet-blog/2021/1/7...
It would be really nice if manufacturers published delta-T vs COP curves.
They use concentrated PV to generate electricity directly, and the cooling water goes into a pond (with insulation on top) that reaches ~90C. They also have a “cool pond” cooled using a heat pump to just above freezing.
The combination of the heat and cold reservoirs are used to run a turbine with a low boiling point fluid.
Obviously with the low temperature differentials it’s not terribly thermally efficient, but the proponents claimed the cost of electricity combined with the storage will make it viable in sunny locations.
Doesn't feel like there is a gross thermodynamic reason it wouldn't work.
4 to 6 liters is 5-8% of that.
5-8% of 1200 watts is 60-96 watts.
If you remember the numbers right, that thing is already roughly in this range. A 100 watt solar panel might fold into a briefcase.
The steam is very hot, and the condensed fresh water is also pretty hot. You can use a heat exchanger to transfer this heat into your 16°C ocean water, preheating it so the sun doesn't have to do all the work.
This wouldn't mean magically getting free energy. Your output fresh water would otherwise be really hot, and that's where the energy comes from.
Obviously there are limits. If you have 100°C fresh water coming out and 16°C ocean water going in, a heat exchanger will at best average their temperatures to 58°C. You could in theory overcome this with a heat pump, but that's pretty elaborate.
(I'm not trying to design the perfect desalination system here. The point is there may be some easy efficiency wins.)
According to GPT4, the numbers come out to 760 J/L for seawater with a salt concentration of 35 g/L. That would mean a limit of around 2 L/s*m2 for full intensity sunlight.
Part of my idea is to use desalinated water to build artificial lakes and rivers. Specifically, a river system along interstate highways, especially in the US west and alongside that infra build highspeed railways. These two things along with interstate highway will allow new towns and cities to flourish which will help with economic activity recouping some of the cost but also solve homelessness, climate goals and even social unrest and instability. It passed my poorly done napkin math.
If hitler had plans to dam the mediterranean and turn the sahara into an fertile land and generate crapton of energy wth are we doing today with all the peace, economic health and insane amounts of technical progress? $300B in subsidizing ev car chargers (car makers should pay for) instead of high speed railways makes me sad.
On a sunny day, total solar energy hitting the top of the atmosphere gives about 1360 watts per square meter (source - top result from google, something something NASA). It takes about 4.184 joules to raise one ml of water by 1 degree celsius. Let's suppose that the "suitcase" proposed in the article has 1 square meter of space (a pretty large suitcase) and is able to capture 100 percent of solar energy (unreasonably effective). Then pure solar energy-wise, there is enough energy to raise 325 ml of water by one degree of celsius per second in a square meter. If the water is, say, at 16 degrees (the temperature of the ocean near me), in would take 84 seconds to bring 325 ml of water to boiling temperature. To transform boiling water to steam, each ml of water takes an additional 2257 joules. In 540 seconds, the 325 ml of water is now converted to steam.
In total, it has taken 624 seconds to steamify 325 ml of water (under very optimistic circumstances). Morally one could do this almost 6 times in an hour, giving approximately 2 litres of water per hour.
(Edit): Looking a bit closer, I see that the article's description of their work suggests that they recapture some of the energy from condensing the steam back into water. I didn't account for this. Even a moderate amount of recapture changes the math very favorably.
This makes me suspect that their claim of 4-6 liters per hour would require a rather large "suitcase"-sized device, but is within the ballpark of reason.