Hacker News Reader

Comment by cconcepts | Hacker News Reader

It surprises me that we haven't worked out a more efficient solar distillation process for desalination. Reverse Osmosis seems so energy intensive at a fundamental level when you consider that the sun naturally distils and desalinates water all day long.

The process described in this article ("Membrane Distillation") seems like a hybridisation of the two techniques and the researchers have found a way to improve efficiency by concentrating heat using lenses. Why can't this concentration of heat be used in straightforward distillation and why does distillation always seem to come in second best in terms of efficiency compared to the use of membranes? Is it because of the large surface area required?

View on HN

Just to fill in details of the other responses. There are multiple concept here: latent heat, specific heat and relative humidity. The latent heat of a substance is the amount of energy required to transform it from one phase to another (in this case from liquid to gas). It varies on the temperature, but basically it's about 2250 - 2500 kJ/kg. However there is only so much water you can evaporate at a particular temperature/pressure. At room temperature it's a lot less than at boiling. If you want to evaporate a lot of liquid, then you need to boil it. This requires you to heat it. The specific heat of water (the amount of energy you need to raise 1 gram of water 1 degree C) is 4.2 kJ/kg. So if we need to raise the water to boiling and then boil it, we will need about 2600 kJ/kg.

If the sun can provide 1 kW/m2, the in 1 hour we can produce a maximum of 1 kWh/m2. That's 3600 kJ. That means we can boil 1 kg (or 1 litre) of water per hour per square meter of sunlight at maximum efficiency. So it's just not feasible.

I welcome fixes to the inevitable errors in the above :-P

You will get the latent heat back when the water condenses again.

This is only true in theory. Thermodynamics says you get it all back, but in the real world it is hard to get it back in a useful form.

That's a principle already present in extant designs from at least the 1960s, see MSF processes (https://blog-en.condorchem.com/evaporation-systems-water-des...). Reverse-osmosis is still far more energy efficient.

The major use of evaporative distillation (based on a few minutes of research, though some general familiarity) is in chemical processing, including but not limited to petroleum distillation. There's well over a century of history and R&D here (the first commercial petroleum distillation occurred in the 1830s, though larger-scale plants date from the 1870s and onward, as kerosene was used for lighting). Optimising for energy use and production volumes, controlability, predictability, safety/risk, etc., has been going on for a while.

Keep in mind that in evaporative desalination you need to both heat the input supply and cool the output, and that processing is rate-limited at both ends. So yes, you could pre-heat the inflow via a condensing jacket, but heat in must equal heat out, and if there's an insufficiency in cooling from native inflow, you'll need an additional cooling working fluid.

Other options include partial vacuum on the salty side, overpressure on the pure side (reducing and increasing boiling and condensation points, respectively), though PV = nRT (Boyle's Law) says that that isn't free either.

And better minds than mine have looked at this for a while.

The issues of input purity and filtration (a major problem with RO desal) is another major factor -- you don't want to just dump raw seawater over your osmosis membrane, but need to pre-filter that through other materials (I'm largely assuming gravel, sand, and possibly clay filters), which will foul with time and require replacement, as well as increase the water-handling energy costs (filtration isn't free).

Reverse osmosis has also seen energy recovery research, e.g., https://scholarcommons.usf.edu/cgi/viewcontent.cgi?article=4...

Possibly interesting reading:

https://duckduckgo.com/?q=energy+recovery+evaporative+desali...

https://scholar.google.com/scholar?q=energy%20recovery%20eva...

Right, but you "get it back" on the condenser surface, not in the fluid that needs to be boiled. At it is exactly the temperature gradient that you need to maintain, so cooling the condenser surface with the to-be-boiled fluid will work only to a point. There are systems that cool the condenser with the water before sending the water to the heater, but the rate at which heat flows is much slower than the rate at which the water must flow.

That's the problem with heat. It is not a source of energy. It is a store of energy, often one that simply needs to be disposed of.

anybody tried solar concentrated boilers ? is it too space consuming ?

Not enough power. The absolute maximum you can get from sun is around 1-1.1kw per square metre.

Thermal desalination plants on other hand, are often as powerful as powerplants.

The main problem with distillation is it requires a phase change, which is extremely energy intensive especially in case of water.

That energy can be recuperated.

It's true that you could build a condenser and pump the steam through the next batch of water, but I doubt you will ever get anywhere near the efficiency you need to make it reasonable. I'd be incredibly happy to see any links that would prove me wrong, though!

The process is called Multiple Effects Distillation https://en.wikipedia.org/wiki/Multiple-effect_distillation

It is a commercially successful process. For example Hitachi have build 8 plants delivering a total of 8,600m3/day https://www.hitachizosen.co.jp/english/products/products011....

That's really cool. Thanks! I should point out, though, that I meant that I don't think you will get efficient enough to make solar power reasonable. You will notice that Hitachi's RO solution has higher power efficiency than the distillation solution. I'd be really curious to see how much power it actually uses, though... Unfortunately it doesn't say in their marketing material.

IANAP but this process will increase the entropy of the energy and result in it being more difficult to utilize

One alternative to reverse osmosis is new technique called shock electrodialysis: https://www.sciencealert.com/scientists-have-figured-out-how...

>Reverse Osmosis seems so energy intensive at a fundamental level when you consider that the sun naturally distils and desalinates water all day long.

Sun does it but it does so over large surface area and the distillation per unit area that naturally happens is much less than what we require.

> why does distillation always seem to come in second best in terms of efficiency compared to the use of membranes?

Offhand I think two or three reasons.

1. You need a temperature differential to drive distillation. Multistage distillation helps somewhat. That temperature differential limits the ultimate efficiency you can achieve.

2. You need to create a phase change for all the water molecules. And then recover as much of the heat as possible. Vs membrane which operates directly on the sodium/chloride ions.

3. Distillation uses really low grade (low temperature) heat. Turning high grade energy sources to low grade heat tends to result in a dead loss. Vs membrane which uses pressure. Pressure is high grade energy.

Reverse osmosis is much more energy efficient than Humidification-Dehumidification.

This isn't true. You can feed the energy gained from dehumidification back into the humidification phase. Last I checked this was able to make the two about equal as far as energy inputs required, but good engineering could perhaps get it better. Reverse osmosis has energy requirements that seem harder to reduce, but of course advances can happen anywhere anytime.

Feel free to fact check the above: I'm not an expert and may have been reading marketing materials that hide something important.

The factoids grade RO at 3.5-6 and MED at 15-20 [kWh/m^3]

RO is also a thermodynamic process, though based on pressure rather than heat differentials. That's also subject to energy recovery, see: https://scholarcommons.usf.edu/cgi/viewcontent.cgi?article=4...

>Why can't this concentration of heat be used in straightforward distillation and why does distillation always seem to come in second best in terms of efficiency compared to the use of membranes? Is it because of the large surface area required?

Energy intensity (and the cost thereof) are a rate limiting factor for thermal desalination:[0]

"Thermal processes are robust, durable and, once in operation, require less maintenance. However, these processes, mainly MSF [Multi-Stage Flash], require large amount of energy for their operation. In fact, up to 50 percent of the operating cost may be attributed to providing energy for the operation of the plant. As a result, many thermal plants are constructed in conjunction with power plants to utilize exhausted steam from the latter. The associated costs of high-energy consumption can have particularly strong restraining effects on small and medium sized plants. However, thermal processes can utilize low grade, low cost source of heat, which would otherwise be lost and discarded to the environment like stack gases, cooling water streams or low pressure exhaust steam. Heat from solar energy can also be used in such processes."

>It surprises me that we haven't worked out a more efficient solar distillation process for desalination.

I'm often surprised at how long it takes to move from discovery to implementation and adoption. The photovoltaic effect was discovered in 1839.[1] Yet 180 years later solar is only 2.4% of the world's installed electricity capacity; transportation and heating/cooling are still overwhelmingly gas/oil.

On an optimistic note, that means humanity probably has tonnes of undeveloped science and engineering that has been discovered but not yet built. It would be infinitely worse if we had already exhausted our collective capacity for invention, yet still had not solved the problems and existential threats we face.

[0] https://pdfs.semanticscholar.org/3f8c/147635ef7dae3050075c2c...

[1] https://en.wikipedia.org/wiki/Edmond_Becquerel

Reply on news.ycombinator.com