Emissive heatsink coatings

[JohnR66]

This is a recreation of an experiment someone performed here a while ago (sorry can't find the link). It is not always possible to convective cool the heatsink well to keep the power LEDs cool and happy. Emissive coatings on the heatsink can play an important roll in getting rid of heat. It is something that many people don't think about when heatsinking their power LEDs.

I attached two power resisters to two aluminum plates of the same size. One was a "control specimen" to check repeatability. I powered them up and let the temperature stabilize for a while and took readings. Next I painted one plate with black latex paint and retested.

Results: (Deg F)
Control: 1st run 136.6, 2nd run 135.5
bare plate 1st run 139.8, 2nd run (painted black) 132.6

Coating the heatsink resulted in 7.2 Deg cooler operation!

Since my heatsinks were vertical with perhaps better convective cooling, the difference was not as dramatic, however, coating with a better blackbody radiator still made quite a difference.

The difference in temp between the test and control pieces is probably due to thermal contact of the resistor to the plate. I used double sided tape (ATG). The large clamps are just to hold them up.

 

[kingofwylietx]

Your experiment does bring up a few questions.

How long did you allow for temps to stabilize?

Did you coat both sides of the aluminum?

What did you use to measure the temperature? Hopefully not an IR thermometer...

 

[SemiMan]

You painted your heat sink with an insulator .... which could quite likely have resulted in a lower temperature measurement as less heat was transferred into the probe.

Semiman

 

[blasterman]

I see this discussion brought up from time, and still don't quite get the science around it. Well, I hear the science, and it still doesn't make sense. I've also been in over 20 corporate sever farms and have yet to note a coated heatsink on tens of millions of dollars of mission critical blade servers, routers, switches, etc.

IBM for instance has spent ghastly sums of money on thermal management research in data centers, but I don't see anything noting how painting heat sinks on a $750,000 iSeries improves thermal dissipation. Same for Cisco, Sun....where are these painted heatsinks? Computer forums are full of geeks that have painted their fans black only to watch their thermal readings climb and not fall.

Assuming for instance this does actually work (and I'm not saying it doesn't), wouldn't the 'black-body' effect work even better if you just dumped the heatsink and emitter in a bucket of black latex paint? Hell, why are we even bothering with a metal heatsink at all? Just paint the back of the emitter and see how well it works.

 

[greg_in_canada]

You painted your heat sink with an insulator .... which could quite likely have resulted in a lower temperature measurement as less heat was transferred into the probe.

Semiman

Agreed. You also need to measure the resistor before and after temperature to show that it is cooler, not just the heatsink.

Greg

 

[saabluster]

I see this discussion brought up from time, and still don't quite get the science around it. Well, I hear the science, and it still doesn't make sense. I've also been in over 20 corporate sever farms and have yet to note a coated heatsink on tens of millions of dollars of mission critical blade servers, routers, switches, etc.

IBM for instance has spent ghastly sums of money on thermal management research in data centers, but I don't see anything noting how painting heat sinks on a $750,000 iSeries improves thermal dissipation. Same for Cisco, Sun....where are these painted heatsinks? Computer forums are full of geeks that have painted their fans black only to watch their thermal readings climb and not fall.

Assuming for instance this does actually work (and I'm not saying it doesn't), wouldn't the 'black-body' effect work even better if you just dumped the heatsink and emitter in a bucket of black latex paint? Hell, why are we even bothering with a metal heatsink at all? Just paint the back of the emitter and see how well it works.

Emissive coatings do indeed help with a heatsink that uses natural convection. The heatsinks designed into computers also come with fans to force air across them. In this case a coating can be a hindrance. If you are designing a heatsink as a radiator it should be designed far different than a normal pin or fin heatsink as those would tend to trap the radiant heat. There is merit in considering a more radiant based heatsink vs a convective heatsink for flashlights since there rarely is much in the way of forced air blowing over a flashlight.

 

[saabluster]

This is a recreation of an experiment someone performed here a while ago (sorry can't find the link).

How about this one? I spent hundreds of hours reading his work and have committed it to memory.

 

[carrot]

How about this one? I spent hundreds of hours reading his work and have committed it to memory.

That's the one I was thinking of as well. Thanks Saabluster.

 

[LEDninja]

Panasonic claims they uses a similar technology in their light bulbs.
http://techon.nikkeibp.co.jp/english/NEWS_EN/20090911/175144/

To enhance luminance efficiency, the company made improvements to the heat dissipation structure. Specifically, to reduce the heat resistance of the LED package and the chassis, it closely attached the substrate mounted with the package to the aluminum chassis and applied alumite treatment to the surface of the chassis. The alumite treatment improved the emissivity by four to five times, according to the company.

Many of the chassis of competitors' products are equipped with fins to enhance radiation performance, but Panasonic's LED light bulbs have a flat surface.

"When light bulbs are attached to lighting equipment, natural convection cooling effects cannot be expected much," Panasonic said. "That's why we chose alumite treatment instead of fins."

The LED light bulbs without fins do not get dusty and tend to be low in weight, according to the company. The cylinder-shaped part of the chassis is formed by press working and as thin as 0.4mm. The masses of the E26 and E17 LED light bulbs are100g and 50g, respectively. In the case of the E26 light bulbs, they are about 40g lighter than competitors' products.

I remember using an electrical cabinet supplier's software to calculate air conditioner requirements for some outdoor cabinets. We (including the suppliers rep) were surprised stainless steel required the most cooling. We were expecting the stainless to reflect the most sun. The manufacturer's tech support pointed out the shiny stainless has the lowest emissive qualities and white painted boxes can get rid of more heat by radiation. NEMA 4X (IPX8) air conditioners are expensive.
Note Panasonic was able to get 4X emissivity without painting the heatsink black. See pictures in the article.

 

[AnAppleSnail]

It makes sense when you consider the airflow around fins in dead air. T(air) - T(fin) will be nearly zero because without airflow the air will heat up and sit. The coatings allow radiation of heat, which I think is based on the temperature of the whole 'outside' - flashlights are unlikely to affect that! Note that my recall of thermodynamics may be spotty.

 

[JohnR66]

Let clear a few things here...

Blasterman: A high surface area heatsink (fins) with forced air over the surface is vastly superior for cooling. This is a simple science experiment. Let's not insinuate, okay?

Latex paint (even white) is a good known blackbody radiator. It is all I had to coat with. It is not an ideal heat sink coating. Anodizing is also good. If the paint were insulating, the heatsink would get warmer and warmer until emission would balance the heat gain.

I coated both sides of the aluminum except were the resistor is mounted. Waited 5-10 minutes until temps stabilized. Did not use IR thermometer since these are dependent on the coating!

Wish I had a high vacuum chamber that would certainly show the effects of IR emission cooling.

Point is if I were to use aluminum bar as a heatsink for my power LEDs where convective cooling was limited, I would absolutely use a high emission coating.

 

[purduephotog]

Let clear a few things here...

Blasterman: A high surface area heatsink (fins) with forced air over the surface is vastly superior for cooling. This is a simple science experiment. Let's not insinuate, okay?

Latex paint (even white) is a good known blackbody radiator. It is all I had to coat with. It is not an ideal heat sink coating. Anodizing is also good. If the paint were insulating, the heatsink would get warmer and warmer until emission would balance the heat gain.

I coated both sides of the aluminum except were the resistor is mounted. Waited 5-10 minutes until temps stabilized. Did not use IR thermometer since these are dependent on the coating!

Wish I had a high vacuum chamber that would certainly show the effects of IR emission cooling.

Point is if I were to use aluminum bar as a heatsink for my power LEDs where convective cooling was limited, I would absolutely use a high emission coating.

'Tis a good experiment.

One thing I've learned is that you have to insulate your emitter, too, so that the heat path is well known.

We use black epoxy to attach thermal probes- not red, not white, not grey, but black. It matters.

So a well insulated heat source with only one path out- the back of the coated plate- in passive radiation conditions should show the results you've seen.

Good work. We tested this once for Chem Engineering- pretty elaborate using mirrored vacuum setups- but the same results.

 

[fyrstormer]

Emissive coatings do indeed help with a heatsink that uses natural convection. The heatsinks designed into computers also come with fans to force air across them. In this case a coating can be a hindrance. If you are designing a heatsink as a radiator it should be designed far different than a normal pin or fin heatsink as those would tend to trap the radiant heat. There is merit in considering a more radiant based heatsink vs a convective heatsink for flashlights since there rarely is much in the way of forced air blowing over a flashlight.

And yet, the brightness and color with which a black-body "glows" depends on its temperature, not on its original color. If this is true in the visible spectrum, I see no reason why it wouldn't also be true in the infra-red spectrum.

What are some "good" emissive coatings, and what are the properties that make them work?

 

[JohnR66]

How about this one? I spent hundreds of hours reading his work and have committed it to memory.

That's the one! His results were amazing. Hard to believe the paint made that much difference. My results were not nearly this drastic probably since I have a larger surface area.

I'd encourage everyone to look at that thread.

 

[saabluster]

And yet, the brightness and color with which a black-body "glows" depends on its temperature, not on its original color. If this is true in the visible spectrum, I see no reason why it wouldn't also be true in the infra-red spectrum.

What are some "good" emissive coatings, and what are the properties that make them work?

Well there is this although it is not sold as an emissive coating. There was some product I was looking for at one time I think called 3M Black Velvet but I never could track any down. Without going and searching for it I do also remember reading that there are also white emissive coatings that match the performance of black so it is not really the color that is key. Although I understand some of what impacts a surface to radiate I don't know the mechanism specifically that makes some so superior to others.

 

[blasterman]

Blasterman: A high surface area heatsink (fins) with forced air over the surface is vastly superior for cooling. This is a simple science experiment. Let's not insinuate, okay?

What I'm 'insinuating' is that big technological companies have spent hundreds of millions of dollars trying to milk a few percents better thermal performance from equipment far more sensitive to heat than anything we screw with here. Not like the LED lighting industry is having issues with heat or anything either and the engineering departments at Phillips, GE, and the entire manufacturing sector of China would love to know about this. Where are the the references to black paint in the passive thermal guides from Bridgelux, Cree, etc. that I've read through?

Also, everthing I've read on the topic from a pure science stanpoint indicates that black-body radiators only have better performance inside a purely radiative evironment - aka, vacuum. While in the presence of air or liquid in a gravity well, convection is the main source of thermal xfer. This can be clearly demonstrated by running a passive heat-sink upside down vs vertically and noting the rather obvious differential in temp.

The term 'black body radiator' also doesn't mean 'painted with black paint'. If refers to an object that is physically a different color than an objects that is shiny. Most paints are polymers, and polymers are insulators. Black paint is a polymer that has a tiny bit of black pigment or carbon in it. Black latex paint is not the same stuff in the Looney tunes cartoons that they paint on walls to make a pefect black hole they walk through. However, this tends to be the assumption with these threads.

So, in order for this to work with paint on a heatsink in your living room the coating could only be a few molecules thick, and ideally on a simple heatsink such as a simple cube or sphere that has little convective performance.

Otherwise, the only real experiment I've ever seen that confirmed the black body effect was two spheres in a vacuum, and the black one had some high tech surface treatment that made it near black and didn't use black colored paint.

 

[AnAppleSnail]

Not looking to start or continue pointed words, but there are two ways to reduce conductive cooling - wrap a heatsink with hot unmoving air, or put it in vacuum (clearly more effective). Which situation will we encounter? Vacuum, or a light wrapped in unmoving air, perhaps in a gloved hand? It seems like the next logical step is to wrap each of these in a tight-fitting sock and compare the temperatures.

 

[Kestrel]

I'd encourage everyone to look at that thread.

:bow: NewBie

 

[bshanahan14rulz]

Great experiment!

Here's some more ideas if you need them.

*same resistor, just one plate, but half of the plate dipped.

*Anodized but not necessarily sealed, instead of painted for the black

 

[fyrstormer]

What I'm 'insinuating' is that big technological companies have spent hundreds of millions of dollars trying to milk a few percents better thermal performance from equipment far more sensitive to heat than anything we screw with here. Not like the LED lighting industry is having issues with heat or anything either and the engineering departments at Phillips, GE, and the entire manufacturing sector of China would love to know about this. Where are the the references to black paint in the passive thermal guides from Bridgelux, Cree, etc. that I've read through?

Also, everthing I've read on the topic from a pure science stanpoint indicates that black-body radiators only have better performance inside a purely radiative evironment - aka, vacuum. While in the presence of air or liquid in a gravity well, convection is the main source of thermal xfer. This can be clearly demonstrated by running a passive heat-sink upside down vs vertically and noting the rather obvious differential in temp.

The term 'black body radiator' also doesn't mean 'painted with black paint'. If refers to an object that is physically a different color than an objects that is shiny. Most paints are polymers, and polymers are insulators. Black paint is a polymer that has a tiny bit of black pigment or carbon in it. Black latex paint is not the same stuff in the Looney tunes cartoons that they paint on walls to make a pefect black hole they walk through. However, this tends to be the assumption with these threads.

So, in order for this to work with paint on a heatsink in your living room the coating could only be a few molecules thick, and ideally on a simple heatsink such as a simple cube or sphere that has little convective performance.

Otherwise, the only real experiment I've ever seen that confirmed the black body effect was two spheres in a vacuum, and the black one had some high tech surface treatment that made it near black and didn't use black colored paint.

Not quite. "Black body radiator" means a solid object that doesn't emit radiation unless heated. A "black body radiator" can be any color with any degree of internal conductivity and surface reflectivity, though for experimental purposes the "black body radiators" they use are actually black, so as to minimize the amount of reflected radiation that would skew emission measurements. Anyone who's ever had the unfortunate experience of touching a glowing-hot piece of metal that was bathed in bright light, thus obscuring its emissive radiation, knows how reflected radiation can screw up such measurements.

You yourself are a (non-ideal) black body radiator, hence why you are visible on a far-infrared camera even in complete darkness. You're just not hot enough to emit radiation in the visible spectrum, and you reflect most of the external radiation impacting you, which I'm sure we all agree are good things.

So anyway, I guess there really are coatings that can cause the surfaces of otherwise highly-conductive objects to glow brighter in the infrared spectrum than they otherwise would given their surface temperature, but the results posted here also suggest that cooling through conduction and convection far outstrip the capacity of everyday flashlight materials to cool themselves through radiation -- at least at temperatures low enough to avoid burning the user. Now, as for avoiding the absorption of heat from the outside environment, the surface treatment is of paramount importance, as anyone who's ever touched something with a black or polished metal surface well knows.