Hacker News new | past | comments | ask | show | jobs | submit login
Passive radiative cooling below ambient airtemperature under direct sun (2014) [pdf] (stanford.edu)
145 points by _Microft 21 days ago | hide | past | favorite | 66 comments

I've been actively working on this technology, goal is making it cheaper and simplify installation. Stanford's a highly reflective surface ~95% combined with stacks layers of silica oxide on a wafer under vacume. The trick too achieving bellow ambient temperature is too reflect nearly all solar energy while emitting strongly in the "atmospheric window". Most silica compounds are well suited as emitters, however the hard part is adding a reflector too the silica and minimising heat transfer from the environment. I've managed to make a meta material paint, reflector and emmiter that achieved bellow ambient temperature, with bulky conventional insulation. as for any effective cooling bellow ambient.

Radiative cooling is just not that strong of heat transfer, what you want too look our for is the research into reflective coatings needrthese systems too function. Review paper: https://www.sciencedirect.com/science/article/pii/S030626191... Shameless plug: https://www.scihouse.space

I'm eager to experiment with a material like this for the application of passive water harvesting in a high humidity environment.

Would you be able to recommend some materials that are perhaps sub-optimal for the task but trivial to assemble from commodity sources to produce this effect?

wouldn't large scale usage of a device like this essentially increase the planetary albedo and help fight climate change? especially if you just skip the "environmental heat transfer" part

There's some numbers here for what it would take.


Basically we need 1W/m^2 of cooling for the earth, so if you could get a radiative cooling device with 100W/m^2 you'd need to cover about 1% of Earth's area

I assume the actual radiative power is greater than the cooling capability, since cooling power = radiative power - heat intake from atmosphere - heat intake from sun

so the math works out even better than it seems...?

Let's say human inhabit roughly 10% of Earth surface. If we cover all rooftops in the world, that ought to make some effect yeah?

Short answer not really https://what-if.xkcd.com/84/

This makes me think: Earth's energy imbalance is around 0.5W/m2, while such a paint sends how much, 40W/m2 through the transparency window?

So we'd only have to paint 1/80 of the Earth. That's ~6.4mln km2, or 2/3 the area of the USA. Still a lot, but not impossible.

I'm sure paint manufacturing scales better than li-ion batteries, and those more than doubled in production volume over the last decade.

All that would do is offset things so that we can pollute more. Not to mention the CO2 and other emissions from such a project. And it would ruin a huge amount of space because of course this would run into “not in my neighborhood”. It’s far better to fix the problem than to paint over it.

Interesting, but even if we solved the temperature problem, we would still have the issue of the acidification of the oceans due to excess CO2. In the end we must remove CO2 from the atmosphere one way or another.

That's more a "we haven't made enough paint to cover a large fraction of the planet" argument than a "what would the thermal ramifications of such an act be" argument. Which I was excited to read about, but alas.

For an interesting toy model related to this, check out Lovelock's Daisyworld simulation:


The argument (and the related Gaia hypothesis) has some important and subtle connections to the facts of climate change. Though even if it's correct, and the biosphere will tend to naturally reassert homeostasis, there's no guarantee we'll enjoy living through it.

You would need a really big roller with a lot of knap too. Although I suppose you could paint most of the midwest and avoid the mountains with a flatter roller.

I'm really ignorant about this whole field.

Your work on passive radiative cooling doesn't sound like biotech or related to biotech but your link https://www.scihouse.space is a biotech lab.

I was just wondering the kind of education/knowledge someone who is working on the cooling technology would have, and was surprised to see a biotech background. So am wondering on the journey to get from biotech to whatever is needed for the cooling tech.

Getting past the mental barrier that engineer's and biologist don't get along. Manipulation of micro structures needed to develop meta materials are on sale that biological systems function at. May have come from a classic engineering of wood and steel, but throw in a foam here and a self-assembling structure there and you can make things no amount of carbide inserts and glue would ever get you.

To compare to an air conditioner. This device has about 40W of cooling per m2.

Apparently in Australia you should size between 80 and 120 W/m2 of air conditioning (I think this is cooling watts rather than power usage watts) - https://www.google.com/amp/s/www.crownpower.com.au/blog/choo...

So that means every square meter of living space needs 2 square meters of radiative cooling (assuming no other passive cooling infrastructure). I suspect you'd see further inefficiencies getting the heat to the passive cooler.

So it's within the same order of magnitude of an ac, but not powerful enough that it would be straightforward to retrofit.

The weakness it has it that it needs a clear sky to work. For overcast muggy days or even a high coverage of cumulus clouds, performance will be absent or degraded.

It works at night, however, and for best results you could maximize insulation and "thermal mass" inside the building and minimize radiative transport through the windows.

The best thing about air conditioning, however, is de-humidification and that is a matter of cooling the air more than you have to and then re-heating it. I live in an 1850 farmhouse and the reason I want a ground source heat pump is that the humidity destroys books and other printed matter. I have inkjet prints curling off the walls and detailed logs of how 3M's best products only work 90% of the time in my applications.

Adhesives are not built for high-humidity environments. My parents live in the tropical rainforest of Panama. One of the challenges is that anything stuck together with adhesive usually comes apart over time. The glass panel on the door of my mom's oven fell off.

As an engineer I don't accept that things have to suck.

If other people think failure is OK I can't do anything about it, but if I have the problem that "Adhesive X does not work in Environment Y" I am going to change the adhesive, change the environment, or not use an adhesive.

I'm having trouble understanding why a clear sky is important. Surfaces radiate based on their temperature and emissivity only, right? So why would it matter what the surface is emitting toward?

Perhaps what I'm missing is that clouds emit some radiative heat back to the surface, whereas a clear sky emits very little, so the net heat loss from the surface under a cloudy sky would be lower.

> Perhaps what I'm missing is that clouds emit some radiative heat back to the surface, whereas a clear sky emits very little, so the net heat loss from the surface under a cloudy sky would be lower.

This is exactly it, yes.

Water vapour re absolves the IR akin too blowing on your sails however it's affect is reduced when it's cloudy can still function just not as well

I am trying to figure out whether one should take into account the cooling power of the surfaces being replaced. The figure of 80-120 W/m2 for air conditioning is presumably based on conventional building materials, which have negative cooling power.

In the paper, the figure of 40 W/m2 seems to be the net cooling power, which is defined in equation 1 as being the power radiated away minus various inflows of heat: radiatively, from the atmosphere; radiatively, from the sun; and by conduction and convection. As far as I can see, these corrections are all for this particular surface, not the surface it might be replacing. These will not, in general, be the same, and, given that this new surface is both highly reflective and vacuum-insulated, I would guess that its values for these properties are lower than the conventional building materials on which the a/c rule-of-thumb is based.

Nevertheless, I doubt that replacing the entire roof with this material would be sufficient cooling, on its own, in the Australian case, and I agree that this would not likely be a straightforward retrofit, to say the least!

This was my thought as well. I have an attic fan and monitor the temperature inside to control it. It regularly gets 120+ F in there on an 80 F day. I have to think the majority of the heating of my house is coming from the attic.

A typical AC cycles its power input. Its not constantly On. One benifit of using this system could be to continuously remove heat from the house and then use an AC on top of it.

There is always a consumption value to free beer.

>> A typical AC cycles its power input.

If it has reached the requested temperature. Like basically all consumer thermostats, it is a bang-bang controller. There is no set on-off cycle. If the AC unit is running at capacity, ie it is properly sized for requirements, it will just be on all the time.

But the temperature goes up and down all the time so there is no "fixed requirement".

If the AC is on all the time it is most likely undersized for the requirement at that time and can't maintain the desired temperature.

In a better-than-consumer setup you will have multiple chillers. Most will just stay on, with one going on-off to handle the variable bit of the load. Starting and stopping electric motors is less efficient tha just keeping them running as much as possible.

Check out interter ACs. Good ones should scale from 20-100% in power or better, so should be able to stay on for most of the time.

Yeah, most central heat pumps installed now are inverter models (also referred to as variable speed). More efficient and more comfortable since you've got a continuous flow of cool air, rather than blasts of cold interspersed with nothing.

There's a lot of potential applications outside of habitat AC in Australia.

For example cooling down stand-alone hardware in the field.

The technology has advanced since then, now white paints with high emissivity in the infrared window are being researched. So if you cover the entire building with that you would get some free, always-on cooling that way.

Opening the windows at night and running fans to equalize the temperature, then putting space blankets over the windows for the day works wonders on hot days in upstate NY.

It's very good as long as two conditions apply:

- the humidity has to be comfortably low

- the outside temperature has to be low enough

Historically, this is usually the case.

In the last few weeks, I've had one or both of those fail to apply on the majority of nights. Dropping to 65F doesn't help when the outside air is also at 99% humidity. If the overnight low is 75F, we're not getting much cooling out of it.

Any system has to deal with time-variable conditions.

I dream of getting a geothermal heat pump for my 1850s farm house which is normally heated with two wood stoves but has a propane backup. (e.g. the kind of compact heater that you see all the time in people's apartments in anime)

At points south the capacity of that kind of system is set by cooling demand but where I live it is set by heating demand. The woodstove could pick up the slack on the coldest days, but that defeats the main selling point of the heat pump which is extreme comfort (e.g. it switches seamlessly from heating to cooling)

- the outside air quality has to be okay

Between pollen count and pollution this is not always the case, and keeping the windows closed and running both the AC and an air filter unfortunately has health benefits for some individuals.

One question that I always wonder when hot days strike:

Given that I only have one portable fan, what is the best setup at night if it is colder outside than inside:

1. Open the windows and put fan so that it blows air out of one window

2. Open the windows and have fan mix the air inside the room

3. Open the windows and put fan on balcony to blow air from outside in.

4. Buy additional fan for ~$15.

Seriously though, it depends. If you have no other way to intake or exhaust air, probably 3, since fans are more effective at blowing than sucking. (IE: 1. would spend some of its power recirculating inside and outside air rather than just pulling inside air out.)

Most likely your bathroom and hopefully stove have exhaust fans, so even better would be to turn one or both of those on, and have the fan blow in a window on the opposite side of the house. It may not even be ideal to open all the windows. You want cool air flowing through the whole house. In an extreme example, if you have the fan blowing in the balcony and an open window right beside the balcony, it could just circulate air there, rather than reaching the rest. Likewise with exhaust, if you have a fan in the bathroom and the bathroom window open. So you'd need to experiment a bit to see what flows air best through the house.

Things also change if you have a central blower.

I had a similar problem in my previous house which didn't have central AC. The insulation in the house was great for about one day. But in an extended heatwave, outside 100F meant eventually inside 90F.

I got decent cooling results with two separate window fans, opposite sides of house, one blowing in one blowing out. Of course, all other windows are closed.

Depending on your window frame style, these types of dual fans fit inside a window frame and don't leak air. They can be set to: both in, 1 in 1 out, both out. So quite flexible. https://www.amazon.com/Bionaire-BW2300-N-Reversible-Airflow-...

A single window fan of the above style can cool a single room quite well, blowing cool air in and blowing hot air out.

As others have mentioned, depending on breeze and temperature, opening all the windows can sometimes work better than multiple fans.

Fans are just no substitute for a real AC unit, which can lower indoor air temp to below outside air temp and can also extract moisture from the inside air.

Another suggestion from your 3 - use an exhaust fan on one end of your home, blowing air from your ceiling out a window(you want to blow the air near the ceiling, it is warmest). Use an intake fan at the lowest elevation possible.

The premise is...cooler air falls, warmer air rises. You want to blow in the low(cool) elevation air, and exhaust the high(warm) air.

I wonder the same thing. If you get a chance, try to test the various configurations!

I would suspect you will also see different results depending if you have multiple windows or just 1. For example, if you have more than 1 window and can seal the opening except the fan, then the fan will move x amount of air in or out and will have similar results.

Simply mixing the air inside the room seems like it's probably the least effective because it will result in very little heat exchange at the windows themselves. However, making the temperature within your house more even may make it more comfortable on the average inside your house.

Matthias tries a few options and evaluates them– https://www.youtube.com/watch?v=1L2ef1CP-yw

> Experiments and anemometer measurements to figure out where to best place a fan to optimally air out the house to cool it down at night.

Nice! His take:

- Blowing out is better than blowing in.

- Fan should be some distance to window otherwise it is ineffective.

Open two windows. Only if needed, put the fan somewhere in between the two windows to facilitate airflow.

That's true, but I don't see why the two couldn't be used in tandem, thus reducing AC power consumption substantially.

Might as well just get solar PV then, that's around 150W per m2.

150W electricity, so you need to run a refrigerator cycle to pump that heat out of your house, which comes at a massive efficiency penalty.

Refrigerators actually have greater than 100% efficiency, often like 300% or so. Because you're just moving the heat, not creating it. Sounds like magic but it's not.

ah true. though that number is coefficient of performance, not efficiency

And an efficient heat pump can move four times as much energy as it consumes. But if some special paint/tile/insulation can reduce your needed cooling capacity by 30% you'll save a lot of money on equipment.

Keep in mind there is nothing preventing you from just angling the device (in its extreme, vertically) and just get an arbitrary amount of radiative surface with a given flat footprint.

(other than of course, it looking unsightly and construction costs)

edit: it would probably help a lot of you angle it such that it is normal to the sun rays, like where i live the sun sweeps from the east to the west, so if you angle the device north or south it would probably work even better.

Are there any resources that catalog historical methods of passive cooling? Many of these methods are space efficient but not cost efficient, and many areas of the developing world (where these issues have the greatest impact) have all the space in the world and very limited access to funds.

I found the question interesting and found a few Wikipedia articles that might help:




Historical passive cooling methods usually involve using shade and high thermal mass like a stone floor - basically simulating a cave.

More sophisticated historical methods include wind catchers used in Persia.

But historically stone houses weren't cheap - poor people lived in straw huts, and most people probably just put up with the heat the best they could cope.

Very cool concept! Has any product come out of this in the 7 years since it was published?

Aaswath Raman's company [1] https://www.skycoolsystems.com

I know of no products but the area of research seems to be active and the article I linked to is getting cited frequently:


Seems like similar tech is now available in paint form? https://www.purdue.edu/newsroom/releases/2021/Q2/the-whitest...

Video from QuantumFracture explaining it: https://youtu.be/wzPdcqrDKzw (You can use subs)

I think I remember seeing a paper where they designed a metamaterial with vertical microstructures that passively lased input heat as narrow band IR within the frequency range that is transparent to the atmosphere. But I can't find it now, perhaps I'm mistaken.

So conceptually, this tech could be used for air-conditioning, during the bright summer day, without consuming electricity?

If so, that's pretty sweet.

Like at what scale does this work? How much is needed to cool, say, a 1000 sq ft apartment in nyc?

This is as insane as the blackbird land yacht which is a vehicle to go directly downwind faster than the wind!

There's nothing insane about the blackbird land yacht, except that so many people seem to think it's insane.

If you just look at the blades as some structure that is moving slower than ground speed through the transmission it becomes much easier to think about.

The surface being pushed by the wind is moving slower than the wind.

Applications are open for YC Winter 2022

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