Hacker News new | past | comments | ask | show | jobs | submit login
Simple, solar-powered water desalination (mit.edu)
501 points by chmaynard 10 days ago | hide | past | web | favorite | 191 comments





Efficiency in desalination is measured by comparing the energy expenditure to the enthalpy of vaporization of water (basically the energy required to distill water by boiling it). The current state of the art reverse osmosis desalination plants use about 3.2 KWh of electricity per cubic meter of fresh water (https://pdfs.semanticscholar.org/d4d7/821d585699719289dddd10...). This technology uses about 173 KWh of solar energy per cubic meter of fresh water. The advantage of this method is lower capital costs and not having to convert solar energy to electricity. For large-scale desalination, however, it is almost certainly more cost-effective to use solar panels, batteries, and large scale reverse osmosis systems. This is still a useful project for making drinking water in remote locations, though.

One fascinating (well, to me) component of modern RO desalination systems is the rotary pressure exchanger. Such a simple device, but so important. It looks like something from the 19th century, but the base patent was issued in 1988.

https://en.wikipedia.org/wiki/Pressure_exchanger


I found this video which seems to animate the concept nicely:

https://www.youtube.com/watch?v=udffed4Pq3g


Bloody Hell I did not know that. I have parts of broken Katadyn PUR-06. I was thinking of rebuilding it from better materials. The pressure recovery is the worst part. I can totally fathom now a hand-held desalinator with only rotating parts running from solar or Li-On batteries.

PUR-06 lasts 300 days, not "indefinitely": https://youtu.be/dNjdgfGgPdE?t=208

I don't know how well this would scale down. Viscosity would work against you.

Thank you, that was fascinating.

Awesome video. One of the better uses of Youtube, and what we should be striving for in terms of "online content".

Awesome. The video made it crystal clear.

That is such an incredibly smart piece of design. A tip of the hat to whom ever thought it up.

Leif J. Hauge

https://www.powershow.com/download/3b604b-ZGQwN/Pressure_Exc...

Read the history, including how the then-current prototype (out of series of improved versions developed over the 12 years it took to get to a market ready design) and all the inventor's personal belongings were lost in the invasion of Kuwait.


So they took the technology of the revolver and applied it to pressure relief. Fascinating stuff.

>One particularly efficient type of pressure exchanger is a rotary pressure exchanger. This device uses a cylindrical rotor with longitudinal ducts parallel to its rotational axis. The rotor spins inside a sleeve between two end covers. Pressure energy is transferred directly from the high pressure stream to the low pressure stream in the ducts of the rotor. Some fluid that remains in the ducts serves as a barrier that inhibits mixing between the streams. This rotational action is similar to that of an old fashioned machine gun firing high pressure bullets and it is continuously refilled with new fluid cartridges. The ducts of the rotor charge and discharge as the pressure transfer process repeats itself.


Thanks for pointing this out. I knew that recovering pressure from the outgoing brine was an important efficiency measure but I assumed it was done using a hydraulic motor coupled directly to a pump arrangement.

very similar in principle to 10 years younger https://en.wikipedia.org/wiki/Pressure_wave_supercharger

That sounds like it would be a great way to eliminate the EGR system on vehicles as well if there was a way to vary the relative speed of the rotor to the crankshaft.

Regarding remote locations- they usually lack water altogether.

A long time ago, as a kid i read about a experiment in a GEO-magazine harvesting water directly from air. The used idea was a hygroscopic chemical (Silicagel) collects the moisture from the air. Sunlight heats the silica gel- which releases the moisture as steam- which then condenses in a destil. It was very little water for a lot of effort, but the complete lack of moving parts and self-containedness of such a system deeply impressed me, being young and with Frank Herberts Dune on my mind.

To drop such a contraption into the deep dessert, where it could keep a plant alive, become a oasis with no basis ..


You can make a solar still without desiccants. Dig a hole somewhere that will get sunlight most of the time, drop a cup in the bottom to catch water, and create a funnel with a plastic sheet. (If you don't have a plastic sheet, branches can work.) There are various ways to optimize it, but a single water trap isn't going to keep a person alive.

But if they're cheap and easy, and a person could make a dozen, that changes things a bit. But still, I think we're still talking about short term survival only. If that's your environment, the local flora & fauna are unlikely to be prolific enough to fill other needs without extensive effort. Unless, you know, you've got a Still suit and can call giant worms and stuff.

The thing is- this could work without intervention - and without the intention of keeping humans alive for years, thus creating a ecosystem over time. Somewhere in the middle of nowhere- watering cacti, then hardened other plants, reclaiming the moisture they loose, until bushes and trees take over these functions.

Its hard describing this fascination with self-contained or only slowly expanding pockets of life.. imagine a glass bubble , filled with small plants, beatles and life, thriving in the midst of a frozzen over wasteland like mars or a dessert like the death valley.

Its like a https://en.wikipedia.org/wiki/Bottle_garden but without the surrounding walls..


To what extent would systems like this extract moisture from the air. siphoning it off from flora/fauna that already make use of it?

There's a newer technology called a metal organic framework that does this at a reasonable rate for the effort.

https://www.youtube.com/watch?v=XPSYzLZ7xKU

https://www.sciencemag.org/news/2019/09/crystalline-nets-har...

http://wahainc.com/


Not so fast.

From the panel to the outlet, you only get out from 10 to 15% of input solar power, while thermal desal can use nearly all of it right away, and it leaves more concentrated brine. Multiple effect desal will probably move the efficiency up a bit. It's nowhere near clear cut.

Also if you have low potential reject heat from power plants, it's idea for desal use.

If you have nuclear as your heat source, then economy-wise thermal desal will beat electric many times over.

I recently read about thermochemical, and electrochmical desalination methods which can be added atop a thermal plant, further increasing thermal desalination efficiency.


>more cost-effective to use solar panels, batteries, and large scale reverse osmosis systems

I can't imagine why you'd use batteries at all. Storing fresh water efficiently is cheap, storing power efficiently is not.


To make twice as much water you may need to either install a set of batteries to run at night, or install twice as much RO hardware to run during the day. Whichever's cheaper.

I made a comment with regards to this a while ago [1]. I'll repeat it here as I think it is relevant:

---

I’ve heard on the grapevine that facilities are now being designed which work at under 100% capacity in order to soak up excess/cheap power.

I believe this is coming about because: 1) the high cost of power and presence of sporadically cheap power makes it economically viable, and 2) designing equipment with lower duty cycles can actually provide substantial cost savings. Ie a facility which only runs 50% of the time is much cheaper to build & run than once which runs 100% of the time.

Sorry I cannot cite sources right now.

[1]: https://news.ycombinator.com/item?id=18929542


It'd be a boon for renewables if energy consuming plant costs were less dominated by investment in equipment.

Imagine doing shift planning at such a site though. Eight days from now, there's a 60% chance of wind in excess of 5m/s between 22.00 and 06.00. But labour costs would be 130% higher than 06.00 to 14.00 on that day. On that morning shift there's a 50% chance of clear skies ...


If only there was some form of programmable electronic computation device that could be used to solve such complex sets of constraints modelled as equations :)

It's not the calculation part that's the problem, it's the investment part (and perhaps also labor part).

I think the savings in planning for low duty cycle probably result from reduced redundancy i.e. installing one pressure pump instead of two. Which naturally means when there is a low chance of excess energy is a good time to work on maintaining the systems rather than just taking one off line and working on the other.

This means that you just staff it at a consistent level and plan the work around the energy levels.


I would be sad to be the worker subject to that shift planning

Just make a water tower to store provide pressure at night.

Because you're likely going to be able to collect more energy than you can immediately use. That makes it effective to do so you can run the desalination even when the sun is down.

The question is if you could build and operate a smaller plant to run at 100% capacity at all times which includes energy storage for less money.

You either have losses in:

* Underutilized solar generation

* Underutilized plant capacity

* Overhead of energy storage

Design your system to minimize that cost. Engineering systems of all sorts are full of this kind of optimization problem: "Should I add a subsystem to recover loss x?" There is always more you can do to recover losses, but each extra system you add suffers diminishing returns a little more.


You actually don't even really need the batteries. Israeli utility scale desal is designed to operate in variable mode, in their case they run about 10% capacity during summer days and 100% during summer nights. Their grid peaks are air conditioning dominated which is why that pattern.

It does mean much larger service water tanks as you need to buffer demand but if you have lots of "free" solar power then it may well be worth it.


You're comparing both systems using watthours of energy expended, so mentioning enthalpy of water vaporization is unnecessary and confusing.

Sorry, I meant to say that the enthalpy of water vaporization is about 630 KWh per cubic meter for context to compare to the energy use of this method and reverse osmosis. I guess I forgot to actually include the number!

Say solar panels cost $2 per watt, 6 liters/hour is about 20 watts of power, is about 40 dollars of solar panels. Plus whatever the reverse osmosis equipment costs.

This system gets approximately 6 liters/hour/sq m... so it makes sense if we can manufacturer it for less than 40 dollars per square meter and space is cheap.


> Say solar panels cost $2 per watt

Not for a large desalination project. You will pay close to the spot price. Which is 30-45 cents.

Your biggest cost will be labor anyway. And the osmosis membrane, which needs to be regularly replaced, unlike the panels.


And, in places without usable water, space tends to be cheap.

I wonder if there is a formula for how cheap a place is based on how easy it is to have usable water and usable crops.

I have had these feelings since I saw the Egypt map showing how much of it is just not usable.


If the capital costs of such a system are lower and they can convert the design into something floating on the sea on continental shelf with tethers mooring it to the floor i see a lot of potential for water, land and electricity (competitive) stressed cities. A case in example is Chennai, India.

I wonder if RO based desalination has any learning rate associated with it and whether a lot of CAPEX costs are related to some proprietary IP. I know that a lot of OPEX is energy costs and hopeful that solar will get even more cheaper with time and even more scale. However if the same can be done with scaling of RO technologies and reducing costs to say ~20 % of existing a great many opportunities will arise.

Really hope it will be lower than 0.1 $ / m3 before 2030.


You seem to know a bit about this: would you have a similar comparison on the capital cost issues, e.g., cost per cubic meter for the at-scale version vs. this local small scale? I mean, that's probably unimportant even if there's a big difference given that remote locations simply wouldn't be practical to have a large plant. But if the costs are even only twice as much per cubic meter, the rapidity of rollout for the small-scale solution could provide rapid relief, especially in some of the areas of the world with acute shortages that need rapid remediation even if it's only a short term solution. (like Cape Town recently)

> For large-scale desalination, however, it is almost certainly more cost-effective to use solar panels, batteries, and large scale reverse osmosis systems.

For large scale desalination, you need to attach tiny magnetic particles (think nano-scale) to the salt, and then extract it with magnets. And as a byproduct, you'll get Lithium.


Is thata dangerous byproduct? Or is the lithium isolated and of usable quality as feed stock for the battery industry?

For a sodium ion in solution it's more a question of pico-scale. And since most sodium salts are quite soluble you'll need some pretty exotic glue to attach that nano-magnetite.

How much space would this thing take up for say, a remote oil rig?

"than 1.5 gallons of fresh drinking water per hour for every square meter of solar collecting area."

Units, people. Units. Be imperial. Be metric. Hey! Be both!


It comes out to about 0.4 liters per square foot, for those of you who are confused.

That means for every square rod of surface area, this unit produces over 11,000 hogsheads of water! Impressive.


A millifortnight turns out to be just over 20 minutes...

How much water can I get in a moon?

If by 'moon' you mean our Moon, it has a volume of 2.1958×10^10 km3. Although, how much of that will hold water, we don't know. Anyway, a jerrycan is probably a lot more convenient ... holds 4.4 imp gals.

I really meant a month, but that was good :)

> 4.4 imp gals

That's 0.062 cubic Flemish ells, for those who were wondering.


To fit my Excel model, I really need to know how many barn-megaparsecs of water this produces per square smoot.

https://en.wikipedia.org/wiki/Smoot

https://www.google.com/search?q=barn%20megaparsec&oq=barn%20...


Matt Groening would be so proud of you right now.

Firkin hell

What's that in bathtubs per football field?

EDITED:

bathtub ≈ 302 liters

football field = 5351 square meters

They achieved 5.78 L/m^2/h

So 5351 m^2/football field * 5.78 L/m^2/h = 30928.78 L/h/football field

30928.78 L/h/football field / 302 L/bathtub = 102.4 bathtubs/football field/hour

h/t to brudgers for the correction about football field surface area!


A football field is 5351m^2. Futbol pitches vary.

The Canadian Football League would like a word with you.

I'm available right after fourth down.

Go Argos?

Oooh, thanks! Updated!

Layman's Explanation Greatest Hits right there. Except I was actually curious so I converted it. Assuming a 30-gal tub, it's about 267 tubs/field-hour.

Hrm - "To start, the standard bathtub will hold roughly around 80 gallons (302 liters) of water. Much smaller bathtubs can only hold around 40 gallons (150 liters) of water, which typically are more suited for smaller children or function more as a shower space."

30 gal seems too small.


well an NFL football field is approximately 5351 m2 so about 8000 gallons. not sure how big your bathtub might be though

im sorry it's 1.25419 gallons per square yard

It also makes the Kessel run in 1.21 JiggaParsecs.

Or, let's see... 5.7 millimeters!

(Per day, one assumes...)

From the paper's abstract:

> 5.78 L/m^2/h

Whoever wrote this PR is responsible for the mix and match units.


I guess it's not really surprising for American engineers. Food and drink are measured in Imperial units. Technical doodads are measured in metric.

You don't really need them to be compatible in this case: even though it appears that you're talking about multiples of units of length, you don't really compare them. You're not going to do the calculation that this is "1.58 micrometers per second" or ".447 feet per day", which is what this is.

I mean, I guess if it helps you visualize a half-foot of water growing on top of the object every day, great. But I feel like "volumes of water" and "areas of solar panel" are quite different units, and it's kinda helpful that the system distinction helps make that clear.


Given the increasing popularity of 500 mL and 1 L units of bottled water, and the ubiquity of the 2 L HDPE beverage bottle, I think it would have been safe to measure the potable water output in liters. About the only things still sold by whole gallons any more are milk, water, iced tea, and lemonade, in the 1 gal jugs, or water in the 5-gal water-cooler jugs.

I'm more confused by the efficiency numbers exceeding 100%, which seems wrong. A theoretical efficiency of 100% would find the solar irradiance energy of 1 m^2 and then find the volume of fresh water per hour that produces an equal amount of energy when its salinity is increased to the mean salinity of seawater (about 3.5%). So if you can get 5000 Wh/m^2/day from insolation, and the energy difference between salt and fresh water is 0.810 W * h/L, 100% efficiency would be a 1 m^2 area producing 6173 L/day of fresh water. You're just dividing the daily energy of sunlight on your panel in Watt-hours by that 0.810, to get L/day.

The units in the article are all wrong anyway. They say "a rate of 5.78 liters per square meter", but there is no time factor mentioned whatsoever. An MIT roof gets mean 4.59 kWh/m^2/day of solar energy, so 100% efficiency would be 4590 * 1000/810 = 5667 L/m^2/day (1 m^3 = 1000 L). If the number given was per day, that's 0.1% efficient. If it's per hour, that's 2.4% efficient.


Servings of beverages in the US are still mostly in imperial measures. A can of soda is 12 ounces; so are most bottles. Beer in restaurants is usually a pint; if not, they'll specify ounces. A Starbucks venti is so called because it's 20 ounces; a tall is defined as 12 ounces. And American recipes still use cups and tablespoons (where the rest of the world has gone almost exclusively to grams, and not volume at all).

The soda industry went to 2 liter bottles back when the country tried to go metric, and it stuck. But most things haven't, including the individual serving sizes. So people never really got the feel for anything smaller than a liter.


Alcoholic beverages sold in containers are almost exclusively in metric. The old standard size was a fifth of a gallon, or 757 ml. When going metric this was reduced to 750 ml which is close enough that nobody would notice. Since it was ever so slightly smaller there was no pushback from the producers. https://en.wikipedia.org/wiki/Fifth_(unit)

Bottled water is often in metric too, even for single serving sizes.


Recipes elsewhere in the five eyes still use cups, tablespoons and teaspoons as well, although they're not the same size everywhere (eg. a metric cup is 250ml).

Containers marked in liters are very common in America, but every once in a while you'll run into some loud-mouthed boomer who starts feigning metric ignorance in a bizarre quasi-patriotic display, often while badmouthing the French. I believe that avoiding this nuisance is the primary reason why imperial volume units are still used at all.

The engineers don't enter into it; they used metric, of course.

I think it is important to use SI units, because it allows you to do comparisons between different approaches to desalination more easily, as I did in my comment https://news.ycombinator.com/item?id=22270160


> Technical doodads are measured in metric.

Not always. I would wager that you are using an electronic device whose circuit board was dimensioned in 'mils', which are not millimeters but thousands of an inch.


> Food and drink are measured in Imperial units.

Maybe if you don't mind losing your global Mars orbiter you don't:

http://edition.cnn.com/TECH/space/9909/30/mars.metric.02/

_everything_ is measured in metric, period.


The next time I'm supplying my orbiter with beverages, I'll keep that in mind.

Don't forget, 10 hours in a day, 100 minutes in a hour and 100 seconds in a minute (1 metric second = .864 imperial seconds)

... and that is why computers traditionally saved 32-bit seconds since the epoch ("Unix time"). These days we have 64-bits typically but the principle is similar. Only use the funny human units when setting stuff that's visible in the UI; in the back we avoid it.

I prefer all measurements of mass converted to stones, and volume converted to the volume of a 1-stone stone from the nearest drystane wall, provided the stone is one-hand wide and high.

How about 5.67812 liters for every 10.7639 square feet?

1.5 gal/(m²s) or 1.58 µm/s

Sort-of related: https://what-if.xkcd.com/11/


But is that a US gallon (3.79L) or an Imperial Gallon (4.55L)?

Do municipal water engineers still think in terms of acre-feet, or has everyone converted to cubic meters?

Mixing cultures makes everyone nicer.

The combination of hydrogen production and desalination using simple techniques are an exciting prospect for this decade.

For architects of 21st century cities, most of these sustainable technologies would benefit from a cleanslate design, rather than trying to compete with builtup 20th century infrastructure.

I think Austrailia is postured to be the biggest beneficiary of all of these "leaps" in technology this century. Most of the desert area is the perfect canvas to start building indoor cities that will be the models/prototypes for the spacestations that the next generation will use to explore urban life outside of our atmosphere. Even building a cruise ship on land would be a good starting point for a 21st century city if they could support several thousand residents with enough freshwater for self-sustaining agriculture and energy to take root.

Even though we won't have saltwater in space, most of these advancements in sustainables fit the roadmap to space more than they're currently reported. Hopefully they'll break ground on new cities using these new resources sometime soon and change the narrative appropriately.


Landships sound like the plot of some dystopian future film; ecosteaders fight off the scavenger bands.

It is already - in the my opinion deeply silly book/movie series(?) Mortal Engines. I haven't read or watched them but saw a video ad and thought that it was the stupidest thing I had ever seen. Anyway it is effectively landships except they are massive steampunk mobile cities which grind up and harvest other sessile settlements and smaller mobile ones. I get the extended metaphor involved of colonalism and enrichment by exploitation of the weak and pillaging instead of their own production but even for YA Dystopian it is just too dumb logistically for me to suspend disbelief.

The books are much better than the movie, but it's not exactly hard sci-fi. They predate the big surge in apocalyptic YA and are much better than they deserve to be. They're not less plausible than "Cities in Flight" or "Marooned in Real-time" or anything with warp drives and laser pistols.

Instead of dystopian, I'm surprised they're not calling them stagnant water communities/cities.

Cities like Detroit and other areas past their prime did not devolve into madmax scenes, but instead are left with a bunch of toxic funk floating in a cesspool.

Definitely not somewhere you'd want to settle down.


Clearly you've not passed through northern New Mexico.

isn't this essentially a 21st century castle?

It's a solar thermal Multi-Effect-Distillation (MED) plant!

I was part of a team that suggested using geothermal heat for desal with MED technology in Perth, Western Australia a bit more than a decade ago!

The advantage over work based (i.e. electrical/Reverse Osmosis) systems is that you do not lose heat via Carnot (or real world, for that matter) efficiency. You utilize more of the original heat, and via the multiple stages ("effects") you can re-use the latent heat of vaporization over and over.

MEDs are really cool gizmos!


>> 1.5 gallons of fresh drinking water per hour for every square meter of solar collecting area.

So a football field could collect around 8,000 gallons of water per hour. Let's say you can get 5 hours of good sunlight per day = 40,000 gallons of water.

>> The team’s demonstration device can achieve an overall efficiency of 385 percent in converting the energy of sunlight into the energy of water evaporation.

>> Theoretically, with more desalination stages and further optimization, such systems could reach overall efficiency levels as high as 700 or 800 percent

So now we're looking at 80,000 gallons of water per day.

>> Unlike some desalination systems, there is no accumulation of salt or concentrated brines to be disposed of. In a free-floating configuration, any salt that accumulates during the day would simply be carried back out at night through the wicking material and back into the seawater.

>> In production, they think a system built to serve the needs of a family might be built for around $100.


If I am reading [1] correctly, NYC needs approx. 1 billion gallons of water per day. So, they would need an array of 12,500 football fields to meet their needs.

[1] https://data.cityofnewyork.us/Environment/Water-Consumption-...


> 1 billion gallons of water per day

* How much of that water need be drinkable?

* How much of that water can be salinated (e.g. for flushing)?

* How much of that water can be partially desalinated (e.g. for showering)?


I like your line of thinking, except now you would need to introduce two independent clean water sources, label things appropriately, retrofit all existing plumbing, etc. also, the salt in the water would be corrosive and require totally different infrastructure.

In the end, I’m not sure if it wouldn’t just be easier to make sure that all the water is potable.


Indeed. If you do only the drinking water, 2L per person per day for 20 million people is only 10566000 gallons or 132 football fields.

And for those who wonder like I did, I counted around fifty of those on that page: https://www.nycgovparks.org/facilities/football

Ok, but if you use those then you don't have anywhere to play. I think more importantly, 132 football fields is about the same as the total floor area of the Pentagon.

Install the solar panels over the football fields as roofing.

If you can afford it, and/or land/actually mounting the panels is the most expensive part, I'd recommend to install photovoltaic panels instead, though, as they are more versatile.

And it would simplify plumbing a lot to put everything in the same place, vs pumping brine up many roofs, and water down again.


This doesn't matter as NYC is built on a river that provides plenty of fresh water every year. It needs to be treated and filtered, but at much less cost than desalination.

Describing the Hudson as drinkable is somewhat ahistorical: https://en.wikipedia.org/wiki/Pollution_of_the_Hudson_River

In any event, NYC gets its water from upstate, not locally: https://en.wikipedia.org/wiki/New_York_City_water_supply_sys...


Still, compared to space needed for desalination, it's not obvious that the current scheme uses less area (I do assume you can still do some things, like hiking, in most of this area though) :

"One of its largest watershed protection programs is the Land Acquisition Program, under which the New York City Department of Environmental Protection (DEP) has purchased or protected, through conservation easement, over 130,000 acres (53,000 ha) since 1997"


Per the original comment about 5 sq mi:

Based on this random article I found from 2015 [1], the average cost per build able sq ft in Manhattan was getting to $1200.

So, if you magically found 5 sq miles in NYC, it’d cost around $167.3B just for the land [2].

I’m guessing you’d want to build it outside the city and pump it in via aqueduct.

Following the article, an array of off-shore desalinated that pumped water back to the city would probably be the most economically viable solution and could probably be built with other kinds of offshore utility systems to try and share costs.

[1] https://therealdeal.com/issues_articles/486631/ [2] https://www.wolframalpha.com/input/?i=%241200+per+square+ft+...


New york city is 468.484 sq mi in size... so we're talking about covering 1 to 2% of it in these panels... I'm pretty sure that much roof space is available.

Of course it probably makes more sense economically to do it offshore or just have a giant field somewhere outside of NYC and pump the water in, but you could find the space inside the city if you really wanted to.


I think you'd lose a bit of efficiency have to pump circulated seawater to the roof tops to not only provide water to desalinate, but also to flush out the accumulated brine overnight.

Those skyscrapers teem with occupants, each with a bladder, each one a source of saline. The spare water and power we use for growing cucumbers in the hydroponic gardens. The waste brine we use for picking those gherkins, which makes the ideal beer-time snack for those occupants as they re-fill their bladders...

Yes, but they get the majority of their water from the Catskills. Water is delivered using the force of gravity.

And assuming a person needs 0.5 gallons (2 liters) of drinkable water per day, we get that 1 square meter provides drinkable water for 15 to 30 people. Not bad!

Of course, once we factor in cleaning uses of unsalted water, the picture is less pretty. Still, not bad.


> So a football field could collect... 40,000 gallons of water.

How many swimming pools of water is that?


40,000 gallons is approximately 150,000 L, while an Olympic pool is in the area of 1,000,000 L.

So around 0.15 Olympic pools — roughly, a football field would fill an Olympic pool in a week.


This is a well-known approach known as multistage flash distillation:

https://en.wikipedia.org/wiki/Multistage_flash_distillation

It's surprising they didn't mention this in the article. (It is, of course, mentioned in the paper, because its authors are decent and honest people.)

> more than 1.5 gallons of fresh drinking water per hour

I don't understand. Are those imperial gallons, US gallons, US dry gallons, or Irish gallons, which are all different sizes? https://en.wikipedia.org/wiki/Gallon#Sizes_of_gallons How many hamburgers is that the same weight as? How many ngogns is that? Or maybe acre-feet, pony shots, or Olympic swimming pools? MIT should be ashamed of their News Office.

The abstract of the paper https://pubs.rsc.org/en/content/articlelanding/2020/ee/c9ee0... does give the result in modern units, and to a much higher precision:

> a production rate of 5.78 L m⁻² h⁻¹ under one-sun illumination

In Dercuano http://canonical.org/~kragen/dercuano I wrote a note about multistage distillation, which I didn't realize was already a known technique until later; see notes/recycling-distillation.html or p. 1776 in the PDF. The basic summary is that modern reverse-osmosis distillation plants require on the order of 7 kJ/ℓ to get freshwater from seawater (which works out to about 50 kJ of sunlight per liter with a cheap 16%-efficient solar panel) while multistage distillation requires a few hundred kJ/ℓ.

The number given by the paper abstract works out to 623 kJ/ℓ if we assume one sun is 1000 W/m² (a common figure for the solar constant at the surface). So if you use those square meters to run photovoltaic panels which you then use to drive a reverse-osmosis machine, you can get dozens of times as much fresh water from the same solar energy. But you need to build or buy a reverse-osmosis machine, and then you need to defoul it, so this device might be a good choice in some circumstances. (You might need to defoul the distillation apparatus as well, though.)


> I don't understand. Are those imperial gallons, US gallons, US dry gallons, or Irish gallons, which are all different sizes?

Imperial and US gallons would both make sense, and you're right it's ambiguous.

Dry gallons would not make sense, since this is a liquid quantity, and Irish gallons are long obsolete.


> Irish gallons are long obsolete.

The word "Irish" in this sentence is unnecessary. Gallons are long obsolete.



> the team’s demonstration device can achieve an overall efficiency of 385 percent in converting the energy of sunlight into the energy of water evaporation.

I need an explanation with pictures for that, because it seems like the author is using ‘efficiency’ incorrectly.


My guess is 385% better than the sun just shining on the water and and evaporating it out in the open

That makes more sense. Maybe 'magnification effect' is better to describe what sounds like possibly a solar-panel-greenhouse system?

It takes 1 watt-hour to produce about 3.5 BTUs.

With a heat pump, for instance, you might be able to produce 5BTUs per watt-hour, because you're extracting heat from the environment (a very large heat source), which would be 140% efficient.

The paragraph you excerpted has your answer:

> The key to the system’s efficiency lies in the way it uses each of the multiple stages to desalinate the water. At each stage, heat released by the previous stage is harnessed instead of wasted. In this way, the team’s demonstration device can achieve an overall efficiency of 385 percent in converting the energy of sunlight into the energy of water evaporation.

You can't drink steam (and even bottling it is futile), so that heat has to go somewhere.

With counterflow systems, like ventilation systems for houses, the inlet and outlet temperatures are pretty close to each other, and the temperature on the 'inside' of the system is either much higher or much lower.

When distillation came up a month or so ago, several people pointed out that modern systems do not operate at ambient air pressure, so the temperature delta may be less than you'd imagine.

[edit: pedants gonna pedant]


I really enjoy artful displays of mental gymnastics, so thank you. A watt is a unit of power, a BTU is a unit of energy, and a watt hour is also a unit of energy. If you can show me how solar watt hours are converted into BTU's at an equivalent rate of 3.85 to 1, I will short sell all the energy stocks using my children as collateral.

A watt hour is actually 3.41 BTUs, not 3.85. A BTU is the energy to raise a pound of water by one degree Frankenstein. (An avoirdupois pound, not a Tower pound, a troy pound, an apothecaries' pound (which happens to be equal to the troy pound), a merchant's pound, an Imperial Standard pound, or a pound sterling or pound of paper, which aren't even units of weight.)

The specific heat of water is one calorie per gram per kelvin, so a BTU works out to one pound, times a calorie per gram, times a degree Frankenstein per kelvin.

A calorie has been defined as having various different values, since water's specific heat varies with temperature (and pressure!) but they're all about 4.18 to 4.19 joules, except for the food "calorie", which is actually a kilocalorie.

Pounds have also been defined as having many different values, even avoirdupois pounds; the values used include 6992 grains, 7000 grains, 7002 grains, and 6999 grains. (Troy grains, not metric grains, which are different.) The currently most popular pound is 7000 grains, but by international agreement it is now defined as 453.59237 grams, previous definitions in terms of the metric system having included 453.59265 grams and 453.59243 grams.

Finally, one degree Frankenstein is precisely defined as 5/9 of a kelvin, although Dr. Fahrenheit's original definition was rather different.

Working all of this out, a BTU turns out to be about 454 g · 4.18 J/g · (5/9)K/K, which is about 1054.3 J. A watt hour is of course 1 W hour · 60 s/min · 60 min/hour = 3600 W s = 3600 J. 3600/1054.3 is about 3.41.

I'll be here all week. Don't forget to tip your servers.


Solar watt-hours are not converted to BTUs at a rate > 1, however, the device effectively uses each watt-hour multiple times, as some of the energy is recovered after each cycle. Needless to say, if the goal were to produce energy, this would be impossible, but much like with heat pumps (which can exceed 100% efficiency by moving heat instead of converting electricity), the device is not producing energy from solar power. The baseline 100% efficiency they are comparing against is that of convertible the water to steam, i.e. the maximum efficiency of a solar desalinator that does not reuse any waste heat.

As a reminder, "Be kind. Don't be snarky." [1]

[1] https://news.ycombinator.com/newsguidelines.html


So that's how this is gonna be.

Good luck.


You mean BTU/hour.

Yeah yeah yeah.

If we're talking about distillation, the civil engineer probably wants to think about rates (liters per minute per $1000), but for a civilian just trying to figure out how this could work in a laboratory situation?

I find quantities easier to comprehend (and relate). Tell me how many AA batteries or days of full sun it would take to convert a big beaker of salt water into a smaller beaker of distilled water.


They are.

I would personally define desalination efficiency of 100% as a perfectly-reversible reaction that establishes an equilibrium between fresh water and oceanic salt water.

Since adding sea salt to fresh water until it has oceanic salinity represents a theoretical maximum of 0.810 Wh/L (a maximally efficient osmotic power plant, situated where a river empties into the ocean, could get about 0.75 Wh/L). So 100% efficiency would be adding 0.810 Wh to one liter of seawater to get one liter of fresh water back. 100% is an unachievable goal, thanks to the laws of thermodynamics.

So to figure your solar desalination efficiency, from a solar panel that receives X Wh/m^2/day of insolation energy, you divide by 0.810 Wh/L to get L/m^2/day. Whatever fresh water you can produce per day, divide by that number to get your efficiency.

An MIT roof gets mean 4.59 kWh/m^2/day of solar energy, so 100% efficiency for them would be 4590/0.810 = 5667 L/m^2/day. By the numbers given, their process is about 2% to 3% efficient (by my definition).

They could be a lot more efficient if they didn't have to overcome the huge heat of vaporization that water has, which is exactly why reverse osmosis is so much more efficient than multistage flash distillation. They are stacking so that the evaporation energy can be recovered from the condensation, which deposits the same amount of energy on the next layer, but it's better off all around to just never evaporate in the first place.


Defined in the paper as

> the solar-to-vapor conversion efficiency, defined as the ratio of total vaporization enthalpy to total solar energy input,

So ya, they are just comparing to a really low baseline...


Examples like this are why I like HN as a community of "engineer and similar minded people". We pick up details like this and call them out.

When I was in Boy Scouts loooong ago, there were instructions and diagrams for making solar stills out of a sheet of plastic, a cup, rocks and dirt.

Same. I wonder how well they worked; I've never seen one made.

It surely would require fairly specific conditions to work well; a wet ground and low humidity.


Just get a large bowl or pail, put an inch of water, or plants, in the bottom. Put a cup in the center. Cover the top with clear plastic film. Weight the center of the plastic. Place it in sunlight.

After a while, condensation will form on the underside of the film, the drops will run downhill to the center, and drop off into the cup.

Voila! Distilled water.

Found it! https://en.wikipedia.org/wiki/Solar_still


Sure, I know the setup. Just never made it nor seen it made.

This is not only about efficiency. As noted in the article, it would reduce the pollution of highly concentrated brine rejected into the ocean.

See for instance this article of the Guardian about a region in Chile where the majority of drinkable water comes from a desalination plant built in 2003: "The salt they pump back in kills everything'" https://www.theguardian.com/cities/2020/jan/02/the-salt-they...


It only does that by virtue of being much less efficient and spread over a larger area than traditional desalination plants. At the end of the day unless you're storing that brine or trucking it off somewhere you have to dump it back in the ocean. That "highly concentrated brine" isn't all that concentrated to begin with as it's less than a 50% increase in solutes coming out of the discharge pipe and gets mixed and diluted in the ocean relatively quickly. That article is just a fisherman blaming an unrelated problem on the closest thing around.

This topic always reminds me of the opening scene[1] from Waterworld where a guy on a raft out in the sea passes his urine through some kind of filter that turns it into drinkable water, and drinks it.

I was a little kid when I watched it and wondered why he would drink his own urine when he was surrounded by such vast amounts of water. With this type of technology, maybe he wouldn't have had to do something so gross! XD

[1] https://www.youtube.com/watch?v=16txiqRoVNQ


Given the agricultural issues in that movie, perhaps the purification device was intended more to collect and concentrate the urea than to purify the water.

But that assumes the scriptwriter would have thought of that, which may be a stretch, given that Earth has insufficient water to cover all landmasses to a depth of 7 km. That would require about 3.6 billion cubic kilometers more water than already exists in the oceans, and there are only about 1.4 billion cubic kilometers of water on Earth to begin with, . Certainly there isn't that much locked up in glacial ice. That is only 26.5 million km^3, which is enough to raise sea level by a maximum of 52 m, far short of the 8 km it would take to cover everything but Mt. Everest.

Anyone who would skip that bar-napkin math was probably only putting that scene in specifically for the gross-out factor. It's just like Mad Max eating the can of dog food. Which is an apples-to-apples comparison, because Waterworld was written specifically to be an ersatz Mad Max.


So the only possible explanation of the cause of the Waterworld apocalypse is earth being hit by millions of giant ice meteors.

IIRC the movie explicitly says "due to global warming" at the start or something to that effect.

But on the other hand, society had collapsed along with most of the stored knowledge of the world, so maybe it was just the people being uneducated and highly insular. They don't even know how long a kilometer is anymore.

There was a point in history a few hundred million years ago where the entire American midwest was a shallow water ocean; we could easily be heading back to that scenario in just a few centuries.


The era of the movie is 2500 AD. To add nearly 3 oceans worth of water in that time frame by cometary bombardment would boil the existing ocean.

The only hard sci-fi explanation is that someone towed Europa to Earth and slowly lowered it down via a cable or pipe. Terraforming by an alien aquatic species. Hyperintelligent space whales.


On a long enough timeline, especially with more energetic atmospheric conditions, isn't it possible for the land to redistribute itself more uniformly from the various forms of erosion to the extent that it's all submerged?

Pretty cool, seems like a modern version of what some people were doing before electricity and solar panels: http://www.solaqua.com/solwatdis1.html

A big question is what do they do with the salt?

Kinda explained in the 9th graf:

> Unlike some desalination systems, there is no accumulation of salt or concentrated brines to be disposed of. In a free-floating configuration, any salt that accumulates during the day would simply be carried back out at night through the wicking material and back into the seawater, according to the researchers.


So the salt is just going back into the ocean. How is that different from other systems, they end up dumping the salt somewhere.

There's no difference. Except that the processing and shipping of ions is passive / much easier. Contrasted with large permeable membrane plants that have to burn power to push all that brine out into the ocean.

But the processing and shipping isn’t the problem with desalination. The problem is that dumping bribe or salt into the ocean can kill local wildlife.

They dump the salt outside the environment. https://www.youtube.com/watch?v=3m5qxZm_JqM

Seriously though, far out to sea away from any landmass it's pretty much a desert. Any salt dumped there would disperse very quickly and not pose a problem to wildlife.


Since the brine is denser than water, a long hose could deliver it to the ocean bottom for just the cost of the hose. There are lakes of dense brine on the sea floor already.

Almost every part of the world ocean has been described as a desert, before people went under and looked. It turns out there's a huge amount of life everywhere you look, even miles down.

But there is also a lot of room, so anything localized, unpleasant, and temporary can be swum away from.


During the day, the salt collects in the capillary material. If it was working 24/7, then it would continuously increase in salinity until you had to do something about it.

However he sun goes down, which means that osmosis works in its favour. The less salty water wicks into the matting, the matting over saturates and leaks back down, equalising the salinity.


From the article:

>> Unlike some desalination systems, there is no accumulation of salt or concentrated brines to be disposed of. In a free-floating configuration, any salt that accumulates during the day would simply be carried back out at night through the wicking material and back into the seawater, according to the researchers.


They'd have enough salt to last forever! >> https://youtu.be/jmzJXVEnmKc

Undercut the existing suppliers of sea salt?

molten reactors

This is pretty cool. It also could solve one of the seasteading problems (water production) in a way that is more scalable. Certainly if they can produce a robust version of this you could pack it into lifeboats so that the first thing you did after abandoning ship would be to toss this overboard and unfold it so that it could be filling/refilling the boat's water tanks.

I am not to sure about it on a metropolitan scale though, that is a lot of surface area per person. One sq kilometer is a pretty big patch for serving up water for about a million people. Call it 800K for losses due to transport.


I guess this an awesome tech for places where there is no clean water, as long as those places are close enough to the shore.

But again, I doubt that any investor will be interested, as they will probably not get good enough ROE with this.

This would probably get the attention of government of some poor countries, or maybe some humanitarian organizations, or even the military.

I would probably be confident that a small portable kit could be marketed to survivalists and it would sell extremely well.



I have a stupid question (well, not to me because I don't know the answer): isn't desalination tech fundamentally similar to alcohol distillation? I know that's energy intensive too, I'm just curious, because I know a little about alcohol distillation and an wondering if the concepts are transferable.

> the team’s demonstration device can achieve an overall efficiency of 385 percent

I don't care how good you think you are, you're not getting 3.85kW per square meter from sunlight.


Would this process also be useful for filtering polluted water?

Several pollutants I have in mind are VOCs and fluorine compounds.

Hopefully there is someone here that is qualified to answer this question.


Probably depends on the vapor pressureS of the VOCs; if less than water, then no.

Is humidity usually high by the ocean? Would it be more efficient to pull water out of the air? Maybe take advantage of low night temps to collect dew?

Why do these articles never have pictures or diagrams?

edit: I just missed them :(


There is a photo and a diagram in the carousel at the top of the page. (Though I agree it's easy to miss.)

It has a carousel at the top with a picture and a diagram.

Nice. When can I buy one at the camping equipment store?

This is great! We need more solar powered projects!

A thunderfoot video waiting to happen

While we wait are there any sites where people actually intelligently discuss this article?

It seems like a 100 year old tech they are trying to optimise. I feel like random optimisations would mean there are no advances to get here.

But it is possible using computer simulations they could do something of value. Have they?


This would make a great DIY project.

Seems good for a seastead.

A lot of lithium batteries are making their way onto both sailboats and yachts. With significant advantages come equal disadvantages, and those disadvantages are the risk of potential battery fires, which only a Class D fire extinguisher can snuff out.

The Dive boat fire in San Diego, where 34-persons perished, is one example.


Class D fire extinguishers are for extinguishing flammable metal fires, involving e.g. lithium, magnesium, titanium, aluminum, and so on. Disposable (primary) lithium batteries contain metallic lithium that can ignite.

Lithium ion batteries contain lithium in an ionic (non-metallic) form. Lithium ion batteries contain flammable electrolytes and polymers, so they can burn, but these are not metal fires calling for a Class D fire extinguisher.


You don't need batteries to run a reverse osmosis desalinator from solar panels; you can run the desalinator when the sun is shining and store the water.

I'm surprised boats are already moving to lithium. I wonder if it would be worthwhile to tow the batteries in a separate buoy behind the boat?


> I'm surprised boats are already moving to lithium.

When the alternative is a bank of lead-acid batteries with effectively ~50% usable capacity and poor charge efficiency, at least for cruisers effectively living at sea off solar power, it's not surprising at all.


WTF! gnaritas's comment has been [dead]ed. This site is turning into a fucking shithole. Here's what he said:

"Lead-acid is still better, it's 1/10 the cost for equivalent usable amp-hours. Most boaters use deep cycle golf cart batteries, they're produced at such scale that they're cheap as hell. Lithium is a waste of money and dangerous to boot. They only make sense if you're extremely space limited, which most bigger boats aren't."

Edit: apparently enough people vouched for the comment, it's back from the dead. But this definitely should not be happening to people like gnaritas who have devoted so much time and effort to making HN better.


The lead-acid batteries still cost a lot less per joule of capacity than the lithium batteries do. They just weigh more. Or maybe charge efficiency is the reason?

Space tends to be at a premium for cruisers, and modern cruisers are increasingly embracing electricity for powering everything.

Charge efficiency is a big one, but it's also just maximizing energy capacity in the available volume.


Lead-acid is still better, it's 1/10 the cost for equivalent usable amp-hours. Most boaters use deep cycle golf cart batteries, they're produced at such scale that they're cheap as hell. Lithium is a waste of money and dangerous to boot. They only make sense if you're extremely space limited, which most bigger boats aren't.

Most marine lithium batteries are LiFePO, not lithium ion. They're far more stable and safe to have on a boat. Compared to lead acid, they have better charge efficiency, deeper cycles, longer lifespans, higher energy density and specific capacity. Furthermore, I've never found lead acid batteries that come close to 1/10th the price...1/4 is more like it.

Or to put it another way, any device with intermittent power only requires one storage reservoir, not multiple. Water jugs are close to the oldest technology we have.

Right. You still need batteries on your boat, just not for desalination.

An interesting aspect of the history of water jugs is that the humans were cultivating them for thousands of years before they started cultivating food crops: https://phys.org/news/2014-02-mystery-bottle-gourd-migration...


Most boats (recreational "cruisers" at least, which is what I am most familiar with) are using LiFePO4 chemistries which are much less likely to start fires. ReLiOn, Victron, BattleBorn, and MasterVolt for example as some of the leading vendors.

Conventional boat fuel burns fiercely, too.



Applications are open for YC Summer 2020

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: