Comment by nkurz | Hacker News Reader

Comment by nkurz | Hacker News Reader

Parent Full thread

The water is condensing on a peltier heat pump.

What makes you so confident that they are using a Peltier? Here's their description of the operation of the device:

During adsorption, air is circulated around the MOF layer and water from air is adsorbed. Passive radiative cooling lowers the MOF layer temperature below the ambient by dissipating thermal radiation to the clear cold sky to increase the effective RH for adsorption. During water production, the OTTI aerogel is stacked on top of the MOF layer to suppress convective heat loss from the solar absorber. The desorbed vapour is condensed on a condenser and the heat of condensation is rejected to the ambient by a heat pipe heat sink.

And here's the more complete description of the condenser:

The condenser of the device was fabricated with a copper plate (4 cm by 4 cm and 0.6 cm thick) attached to a commercial air-cooled heat sink (NH-L9x65, Noctua) to efficiently dissipate the heat from condensation to the ambient.

Are you still sure it's condensing on a Peltier cooler? The paper never mentions "Peltier". If the "condenser" is actually an electrically driven Peltier, this would seem like a fraudulently bad description, justifying retraction of the paper. Is it possible that you are wrong?

Still, I agree that you might be right about the physical limitations of scaling. You seem knowledgeable about the field, and I'd be interested to hear your impression after you read the actual paper: https://www.nature.com/articles/s41467-018-03162-7.

(I am genuinely interested in hearing your opinion about what they are doing, but would strongly suggest less overconfidence and more humility when offering bombastic pronouncements on papers you haven't read.)

> What makes you so confident that they are using a Peltier?

Because you can buy the exact device on Amazon (Noctua heat sink sold separately)[1]. Compare that image to the images shown in this MIT news article[2].

> The paper never mentions "Peltier". If the "condenser" is actually an electrically driven Peltier, this would seem like a fraudulently bad description, justifying retraction of the paper.

I would agree.

> The condenser of the device was fabricated with a copper plate (4 cm by 4 cm and 0.6 cm thick) attached to a commercial air-cooled heat sink (NH-L9x65, Noctua) to efficiently dissipate the heat from condensation to the ambient.

This describes a Peltier device exactly. Pass a current through it and one side will get hot and one side will get cold. Hence the heat sink. You need someway to dissipate that electrical energy.

Ask yourself, if the system works passively (i.e. ambient temperature), why do they need a heat sink? Would attaching a heat sink to my shed cool it off? I think the answer is clear.

Ignoring everything I said, to condense water from air you need two things. Air with water in it and a temperature differential. How is the differential generated? According to the paper, a condenser. How does every condenser generate a temperature gradient? Electrical current.

1. https://www.amazon.com/TEC1-12706-Thermoelectric-Peltier-Coo... 2. http://news.mit.edu/2017/MOF-device-harvests-fresh-water-fro...

As long as you are clear that you are accusing the authors of outright fraud, I appreciate and applaud your logic!

I think you are wrong, though.

Ask yourself, if the system works passively (i.e. ambient temperature), why do they need a heat sink?

Because as you say, they need a cooler surface than the ambient for the water to condense on. Ambient temperature in this case refers to the temperature inside the solar chamber containing the saturated sorbent and insulated by the translucent aerogel. In the "legit" view, the heatsink is cooled relative to this ambient by being thermally coupled to the cooler outside air. In the paper, Figure 3 shows the temperature differential as about 40C: https://www.nature.com/articles/s41467-018-03162-7/figures/3

Would attaching a heat sink to my shed cool it off?

Well, if the inside of the shed has been heated by the sun so that it is warmer than the outside air, then yes. The heatpipes are effectively windows for heat to escape from the relative hot interior to the relatively cool exterior. Isn't this exactly how a passive heatpipe cooler like the Noctua works when installed as designed to cool a CPU? Some airflow over the radiating fins doesn't hurt, but convection takes care of this if the surface is large enough.

I'm sorry I chose my words very poorly.

I said...

> Would attaching a heat sink to my shed cool it off?

What I meant to say...

> Would attaching a heat sink to my shed cool it below the ambient temperature?

To which the answer is obviously no.

A brief lesson on humidity to wrap up the discussion. The numbers I'm giving are bogus but the logic is sound.

Suppose you have a vessel of air at a temperature of 25C and a relative humidity of 90%. If I increase or decrease the temperature have I added or subtracted water from the system? No the water is constant. What will change in the relative humidity of the system.

Say I heat the vessel to 40C. The relative humidity will drop from 90% to 50%. Why? The air can hold more water. But say I reverse course and drop the temperature to 10C. Then the water begins to condense because the air can not physically hold the water any longer.

[/end-bogus-numbers]

So if we take MIT's device and try to extract water from the air we know two things. We need warm, humid air and we need to cool it until it condenses.

A heat sink will not cool the air below ambient. This is required for condensation. No configuration of heat sink + device will decrease the temperature unless we put some "work" into the system. That "work" will be electrical energy of some sort The source is irrelevant. It must be external to the system and it must have a cooling effect. We absolutely need some sort of peltier device (or other refrigeration technology). There is no way around it!

But to further illustrate the absurdity of all of this lets ask the following question. How much water can we reasonably expect to extract? After all 1 liter of air doesn't translate to 1 liter of water.

At 25C and 100% humidity air can hold about 0.02ml of water. Meaning for every one liter of water you need 50,000 liters of air. And suddenly you being to realize the magnitude of your problem. To extract 1 liter of water per day you would have to ram one liter of air through the system every second. And that's in an environment with 100% humidity! It's raining for god's sake! In the desert the amount of air needed will easily triple.

> Would attaching a heat sink to my shed cool it below the ambient temperature?

To which the answer is obviously no.

Except it's "yes" if you define "ambient temperature" as the temperature inside the shed on a sunny day, and if you put the radiating fins of the heatsink in the cooler outside air, which is the equivalent of what the paper does! You can criticize that this is a poor definition of ambient, but it is the one they use.

No configuration of heat sink + device will decrease the temperature unless we put some "work" into the system. That "work" will be electrical energy of some sort The source is irrelevant. It must be external to the system and it must have a cooling effect. We absolutely need some sort of peltier device (or other refrigeration technology). There is no way around it!

So when I take a shower and then and then see water condensing on the inside of my single pane bathroom window on a cold night, it means there's got to be a hidden Peltier cooler in there somewhere? Or in a more exact comparison, if I have an outdoor greenhouse on a cold but sunny day, and I see water condensing on the inside of the glazing, there's got to be a refrigerator making it happen? No, there just needs to be a temperature differential, and the heat sink acts like a cold window.

At 25C and 100% humidity air can hold about 0.02ml of water. Meaning for every one liter of water you need 50,000 liters of air.

By contrast, this is an appropriate criticism! Yeah, to collect a liter (1 kg) of water at 25C and 50% humidity, you'd need to extract all the water from 100,000 liters of air. And since you aren't going to get all the water out, you it probably needs to process several times that amount. Is this absurd? It probably depends on how much water you need. An Olympic Swimming Pool is 50m x 25m x 2m, so that amount air would give you something between 1 and 25 liters of water depending on your efficiency of extraction.

How do you get this much air over your collector? It looks like a medium bathroom fan (70 cubic feet per minute) seems to move a little over 100,000 liters per hour, so it could move this amount of air in a day. Alternatively, if you wanted to do it without power, you'd probably get more than this amount of air flow with a gentle breeze. Whether it's worth it would depend on how thirsty you are in that desert and what your other options are, but it's still a neat technique!

You're misunderstanding the point.

The world is ambient. Your device is not. You are trying to extract water from the world so your device needs to be lower than ambient.

> So when I take a shower and then and then see water condensing on the inside of my single pane bathroom window on a cold night, it means there's got to be a hidden Peltier cooler in there somewhere? Or in a more exact comparison, if I have an outdoor greenhouse on a cold but sunny day, and I see water condensing on the inside of the glazing, there's got to be a refrigerator making it happen? No, there just needs to be a temperature differential, and the heat sink acts like a cold window.

The hot, moist air from your shower should be considered ambient. Your window is colder than that ambient temperature. Hence the water condenses.

Let me explain further because this is clearly a sticking point.

The air your shower is generating is roughly 40C at 100% humidity. You don't have to put in any work to condense this water on the window because you already put work in to heat it.

"Ah ha!" you say. "They're heating the water vapor up inside the container! Just like my shower!" Except no the two are not equivalent. Heating up 25C air at 100% humidity to 40C would decrease its humidity. I.e. no condensation. Cooling 40C air at 100% humidity to 25C would result in condensation (your shower)! It is a one way process. The source of the water vapor MUST be hotter than the condenser! When your source is the atmosphere that means your condenser has to have a lower temperature than the exterior conditions.

> Except it's "yes" if you define "ambient temperature" as the temperature inside the shed on a sunny day

The temperature is higher in the shed but the humidity is lower! Lowering the temperature of the shed to the outside temperature will not produce condensation!

Edit: Sidenote. Greenhouses generate their own water vapor (transpiration). Their absolute humidity will be higher than the outside.

Edit2: I also want to emphasize that if you're adding water to the system you're doing it wrong. We're trying to harvest water not supply it. Showers and greenhouses are not apt comparisons.

I think you are thinking about this device as if it operates continuously. My understanding is that it 'consumes' the temperature gradient created by the day/night cycle.

Could mean that this would be ineffective in the tropics where the temperature between day and night varies less. And presumably the device may 'stop working' part way through the afternoon if the previous night was too warm relative to daytime temperature.

> A heat sink will not cool the air below ambient. This is required for condensation. No configuration of heat sink + device will decrease the temperature unless we put some "work" into the system.

A heat sink surrounded with foam with an opening left that looks down the focus of a satellite dish that has been carefully covered in tinfoil will get colder than ambient air temperature, if the dish is pointed at a clear sky.

I'm not sure if you looked at the paper, but they take advantage of this phenomenon to lower the nighttime moisture collection temperature, and say it gives about a 3C benefit even without a satellite dish:

Operation in such arid regions also opens an interesting avenue for increasing water harvesting output with passive radiative cooling by leveraging the typically clear sky. The clear night sky and low vapour content in the atmosphere enables dissipation of long-wavelength (infrared) thermal radiation from the device to the cold sky to cool it below its ambient temperature. By facing the device to the sky during adsorption, a ~3 K temperature drop was achieved, which corresponds to an increase in 5–7% RH experienced by the adsorbent. This passive cooling can lead to opportunities to utilise other adsorbents that have their adsorption steps located beyond the typical levels of RH in specific regions.

https://www.nature.com/articles/s41467-018-03162-7

Why does the clearness of the sky matter?

Wouldn't cloud cover just cause the heat energy to be absorbed by the clouds or reflected back to a much larger area of the ground (so only a small percentage would return to the device itself)?

I agree that the common explanation is confusing. If I understand it right, it's not the transparency of the sky to the outgoing radiation that matters, but the absence of a (relatiely) high temperature emitter like a cloud. The two are closely tied (if there is no absorber, then there is no emitter) but the "clear sky" explanation mixes up cause and effect.

Rather than emphasizing the transparency, a clearer explanation would emphasize the absence of a counterbalancing warm mass whose emissions can be absorbed. Objects are always losing energy by black body radiation, but normally are gaining thermal energy from their similarly temperatured environment. One a clear night, there is a greater imbalance.

Fascinating, thank you for the explanation

I understand why this works, but still feels like magic. Very cool.

> What makes you so confident that they are using a Peltier?

The article contains a link to an older one and says the device is based on that: http://news.mit.edu/2017/MOF-device-harvests-fresh-water-fro... . Notice the heat sink? Notice quotes like:

In fact, the system doesn’t even require sunlight — all it needs is some source of heat, which could even be a wood fire. “There are a lot of places where there is biomass available to burn and where water is scarce,” Rao says.

That sure sounds and looks like a peltier device to me. MIT is gaining a reputation for crackpot stuff like this.

Yes, there are previous devices that use a Peltier cooler to condense water directly out of the air. The article you link mentions these as a contrast to what this device is doing: "Another method of obtaining water in dry regions is called dew harvesting, in which a surface is chilled so that water will condense on it, as it does on the outside of a cold glass on a hot summer day, but it “is extremely energy intensive” to keep the surface cool, she says, and even then the method may not work at a relative humidity lower than about 50 percent. The new system does not have these limitations."

This approach uses a "sorbent" that adsorbs water at night, and then then uses sunlight to generate heat to drive the water out of the sorbent during the day. It does not use a thermoelectric cooler: 'The new system, by contrast, is “completely passive — all you need is sunlight,” with no need for an outside energy supply and no moving parts.' Rather than requiring active cooling, this approach requires heat to release free the water from the sorbent. The passive heat sink is used to help capture the water after the heat from the sunlight forces it out of the sorbent: "The desorbed vapour is condensed on a condenser and the heat of condensation is rejected to the ambient by a heat pipe heat sink."

So as best as I can tell, your theory that the paper (which you haven't read) is simply lying about what they are doing? Not impossible, but I think it would require some greater level of evidence the "Notice the heat sink". And while we're at it, how does "biomass available to burn" imply that they are using a thermoelectric aka Peltier cooler? https://en.wikipedia.org/wiki/Thermoelectric_effect

(Yes, I agree that it would be nice if the press release would actually link to the paper that correctly describes the apparatus, but its regrettable failure to do so does not give license to make up your own details as to how it works.)

The technique described sounds similar to how desiccant dehumidifiers work, where you capture moisture in a desiccant then remove the moisture by heating the desiccant. The energy required is similar to the chiller based technology due to the laws of thermodynamics.

I have a one that has a descant drum, and it slowly rotates as it blows air through it. A small section of the drum is heated to release water into a separate air stream and that air is then ran through a radiator to cool it, releasing water into a storage container. It costs $0.5-1/L to generate water with it, since the heat comes from electricity. Heat could easily come from solar, leaving ~20w for a fan.

Reply on news.ycombinator.com