heat sink questions 4

[limct]

hi folks,
I read an article about heat sink, it says that;
1) heat sink with painted surface will have better thermal emissivity than an unpainted one
2) and one with brighter color (eg. white) will have about 3% higher in thermal emissivity than a black coloured one.
3) anodizing will decrease the surface thermal resistance.

Can anyone explain the theory behind these?

Thanks a lot.

Best regards,
ct

 

[btrueblood]

The statements given are generally true for aluminum heat sinks. My values below come from Holman's Heat Transfer, 6th ed., 1986.

1. In general, bare polished aluminum has an emissivity of .04, rising to 0.2-0.3 when heavily oxidized. Glossy white or black paints will have emissivities of 0.8-.95. But, unless your heat sink is hot enough to glow, or operating in a vacuum, conduction and convection dominate the heat transfer, and the emissivity is moot. For certain heat sinks, adding a coat of paint will dramatically worsen the heat transfer, as the paint fills the fine details and reduces the surface area exposed to air flow.

2. Not generally true, typical flat black paint has an emissivity of 0.96-0.98; white or black glossy enamel 0.8-0.95. There are some specialty white paints with high emissivity, but they cost more. White has an advantage that it tends to have low solar absorptivity, i.e. doesn't absorb ambient solar radiation, thus reducing the total heat load. The discussion of radiant vs. convective heat transfer in 1) above applies here too.

3. Anodizing decreases thermal resistance? Hmm. If your aluminum is heavily oxidized, or will grow an oxide layer over time (due to corrosion), then that oxide layer acts as an insulator. Anodizing, and its associated coloring/clearcoat/sealer will "fill in" the porous oxide, and possibly improve its thermal conductance (be a better conductor of heat), and also prevent future oxidation. There are commercially available heat sink compounds, gap fillers, pads, etc. that might do a better job of conducting heat; most of these probably call for having a clean, bare metal surface to work optimally. Since the compounds generally do a good job of sealing out air, the chance of a growing oxide layer is much reduced.

 

[Heckler]

Here is a link that will give you some information on thermal analysis.

http://www.coolingzone.com/Content/Library/Papers/index.html

Best Regards,

Heckler
Sr. Mechanical Engineer
SW2005 SP 5.0 & Pro/E 2001
Dell Precision 370
P4 3.6 GHz, 1GB RAM
XP Pro SP2.0
NVIDIA Quadro FX 1400
o
_`\(,_
(_)/ (_)

Never argue with an idiot. They'll bring you down to their level and beat you with experience every time.

 

[ko99]

The article sounds correct but keep in mind some practical considerations.

1. As btrueblood said, painted surfaces radiate much better than unpainted, and in general a dull surface radiates much better than a glossy one.

However radiation CAN make a big difference in natural convection (no fan) even without glowing heat or a vacuum. Here's a tool for estimating heat sink radiation

http://www.heatsinkdesigner.com

2. I suspect the 3% higher radiation mentioned in the article was due more to flatness than color. Color doesn't matter for radiation unless the heat sink temperature is about 5800K! It can matter for absorption, of course, if the sink is exposed to sunlight.

3. Yes, an anodized surface is a thermal (and electrical) insulator BUT it's typically so thin the thermal effect is negligible and, as btrue said, it prevents corrosion which can be worse.

Anodizing also helps radiation. Emissivity is about 0.8, not quite as high as many paints, but much better than bare or glossy surfaces.

I also agree with btrue that you should use a thermal interface material between the heat source and the heat sink, but I don't think you have to worry if that surface of the sink is anodized.

ko (

http://www.ecooling.biz)

 

[kenvlach]

limct,
Please give a link to the article, or give some pertinent information, with quotation marks & reference.

1) True, polished aluminum has a very low heat emissivity of 0.04, and the proper paint can raise this above 0.9. But, emissivity is important even at typical electronics temperatures (see 3. below). And, test results on coated aluminum sheet intended for plasma display enclosures had results comparable to black anodized aluminum, or about 10^oC cooler than bare aluminum. Notice the very thin chemical conversion coating on both sides of the sheet; primer for the paint & protects against the aluminum from forming its native oxide which thickens when heated.

http://www.furukawa.co.jp/review/fr026/fr26_13.pdf

2) Agree with btrueblood on this one. Flat black is best for heat sinks.
For exterior applications where weathering resistance and high solar reflectance (albedo) are desired in addition to high IR emissivity, studies by Oak Ridge National Laboratory, Lawrence Berkeley National Laboratory, and the Florida Solar Energy Center found that white paint, either acrylic or latex, with titanium dioxide (or tin dioxide) pigment, was best.

3) "anodizing will decrease the surface thermal resistance" -- perhaps taking something out of context? Anodic coatings have a thermal resistance which increases with thickness.
-- T. R. Ogden et al., Materials Letters, vol. 5, pp. 83-87 (1987).
The effective thermal conductivity increases considerably with contact pressure.
-- G.P. Peterson & L.S. Fletcher, J. of Heat Transfer, vol. 112, 579-585 (1990).

Additional studies on thermal properties of anodize reported in The Surface Treatment and Finishing of Aluminum and Its Alloys, 6th Edn., pp. 1092-1102 (2001) show that significantly more heat is lost from anodized than from bare aluminum at 30^oC, and 68-79% more (varied with anodize thickness & color) more at 90^oC. Recent (1996, 1997) studies show that IR emissivity of anodized Al 6061 reaches 0.70 (clear) and 0.72 (black) at an anodize thickness of 6 microns, and 0.75 and 0.80, respectively, at 11 microns thickness. Max. 0.84 & 0.90, respectively, at greater thicknesses, but with greater thermal resistance.

> 1 billion anodized 6000-series Al heat sinks can't be wrong!

 

[GregLocock]

OK, I only know enough to be dangerous. 6k aluminum is great stuff, largely because it isn't soft. It isn't soft because it has been alloyed. These alloys will upset its crystallography. I would have thought that this would increase its electrical resistance somewhat. Electrical resistance and thermal conductivity are very closely related for metals.

So, why make heatsinks from 6k as opposed to a pure aluminum?

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[zekeman]

quote
"OK, I only know enough to be dangerous. 6k aluminum is great stuff, largely because it isn't soft. It isn't soft because it has been alloyed. These alloys will upset its crystallography. I would have thought that this would increase its electrical resistance somewhat. Electrical resistance and thermal conductivity are very closely related for metals.

So, why make heatsinks from 6k as opposed to a pure aluminum?

Cheers"

Because, the alloying effect is small so that pure aluminum and its alloys have virtually the same thermal conductivity.

 

[zekeman]

Quote from blueblood
". In general, bare polished aluminum has an emissivity of .04, rising to 0.2-0.3 when heavily oxidized. Glossy white or black paints will have emissivities of 0.8-.95. But, unless your heat sink is hot enough to glow, or operating in a vacuum, conduction and convection dominate the heat transfer, and the emissivity is moot. For certain heat sinks, adding a coat of paint will dramatically worsen the heat transfer, as the paint fills the fine details and reduces the surface area exposed to air flow."
--------------------------------------------
I strongly disagree with the statement referring to the dominance of convection unless the heat sink glows. My research shows that a black body with a surface temperature of 300 deg F (hardly glowing) is twice as effective as natural convection in transferring heat. From this, it would seem that radiation heat transfer is the dominant process for heat sinks.

 

[zekeman]

Correction,
I just reread the quote and and withdraw the previous comments but I might add that the statement "hot enough to glow or operating in a vacuum , conduction and radiation dominate..."
is ridiculous. What other means of heat transfer is there besides radiation and convection and what has vacuum or glowing got to do with the argument?
The fact is that radiation trumps convection in any heat sink application as I pointed out in my previous withdrawn post.

 

[CoryPad]

To answer Greg's question, and to partially agree with zekeman and partially refute him:
[tt]
Alloy Yield Strength Thermal Conductivity
----- -------------- --------------------
1050-O 28 MPa 231 W/(m K)
6063-T6 214 MPa 201 W/(m K)
6061-T6 255 MPa 167 W/(m K)
[/tt]
Alloying can have a huge effect on thermal conductivity (and strength, and corrosion resistance, etc.). However, alloy 6063 (Al+0.7%Mg+0.4%Si) has an order of magnitude higher strength than alloy 1050 (99.5% Al) while retaining 87% of its thermal conductivity. This has been found to be an acceptable trade-off between these two properties. However, if one uses the popular 6061 alloy (Al+1%Mg+0.6%Si+0.3%Cu+0.2%Cr), the decrease in thermal conductivity become much higher.

As far as why do vacuum & glowing have to do with the argument, obviously no one would say that convection is an operating mechanism under vacuum conditions, while everyone would agree that glowing objects radiate significantly more than non-glowing objects.

Regards,

Cory

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[kenvlach]

From the Aluminum Extruders Council (AEC) FAQ's:
"I'm in the market for some extruded aluminum heat sinks. Which alloy is best for this type of application? The purpose of any heat sink material is the dissipation of heat. Although all aluminum is capable of this, the specific alloy 6063 works best, when all aspects of aluminum in this application are considered. This alloy finishes well, and is machinable to most user needs. It's one of the most cost-effective alloys for this purpose. This alloy's characteristics lend themselves well to the sharp corners and tight spaces of some heat sink designs.

Can an extrusion provide enough heat dissipation to cool my electrical components? With only copper, silver and gold having better thermal conductivity, aluminum historically provides the most cost-effective solution while the extrusion can be designed to allow the maximum cooling area where it's most needed. Designing in serrated surfaces greatly increases the cooling area."

http://www.aec.org/techinfo/faqprop.html

The USA tends to use the slightly stronger Al 6061-T6 rather than 6063.

A Google search suggests that China mostly uses anodized Al 6060, 6061, 6063 and K7 for heat sinks.

From Aluminum, Vol. II, Design and Application, pages 551-2 & 555 (1967):
"The extrudability of aluminum permits...Configurations ideal for heat dissipation can be produced economically (Fig. 12). Extruded aluminum, usually of 6xxx series alloys, can be color anodized to increase the emissivity of the radiating surfaces..."

From 'Extrusion of AlMgSi Alloys,' O. Reiso, MATERIALS FORUM VOL.28, pages 32-46 (2004):
"The amounts of the different alloys used for extrusion purposes are shown in Figure 2 [2].* Clearly the dominating alloy system for extrusion of aluminium is the 6000 alloy system, i.e. AlMgSi alloys. More than 90% of the total extruded volume is estimated to be made from this alloy system." [for Western Europe, 2000]
*Reference [2] is a Hydro Aluminum document.

http://www.materialsaustralia.com/Materials_Forum/Vol28/INV 4.pdf

Heat Sink Selection, Thermal Management of Electronics, (PowerPoint presentation) San José State University,

http://www.engr.sjsu.edu/ndejong/ME 146 files/Heat Sink.ppt

"Al alloys used [for casting] have lower conductivity than Al alloys used for extrusions"
'Aluminum (6063 or 6061) is most common..."
"Ideal conditions mean:...The heat sink is painted or anodized black....Thermal interface resistance minimized using grease or pad and by making surfaces as flat as possible"

Some reference on heat sink design (optimization) are given in the final slide.

Also, NASA uses black anodized Al for radiators in space, sometimes even grit-blasted (pre-anodize) for greater surface area. Anodize (clear, gold & black) is mostly used for the International Space Station truss structure, since its IR emissivity keeps it from heating and expanding as much as bare Al in sunlight.

 

[ko99]

My 2 cents...
1. Relative importance of radiation
This is very complicated to predict. It depends on conduction and convective contributions as well as emissivity, etc. A very general rule of thumb is that radiation is significant in natural convection and negligible in forced convection.

2. Color and emissivity
kenvlatch wrote "2) Agree with btrueblood on this one. Flat black is best for heat sinks."

Sure, a particular flat black paint may have tested higher in some published data, but the high emissivity is due to it's dullness not it's color. I stand by my earlier statement that color only matters for absorption when exposed to the sun (and here you want white to reflect heat). Color only effects radiation if the surface is extremely hot (5800K)

3. Anodizing
I think we're all on the same page on this one. Anodizing helps radiation and adds a (usually negligible) thin layer of thermal insulation.

Regarding Aluminum
In my experience, 6063 is the most popular alloy for forged and extruded aluminum heat sinks.

ko (

http://www.ecooling.biz)

 

[ScottyUK]

Just a view from the electrical world where we use a lot of heatsinks: convection is very important because we usually use forced air cooling in order to keep the heatsink size, weight, and cost within sensible boundaries. Most of the force-cooled heatsinks have poor free air heat dissipation in the event of fan failure, so that's a pretty convincing argument that convection is the primary cooling process.

There is a fair bit of information about commercially available heatsinks on

http://www.aavidthermalloy.com/

----------------------------------

Sometimes I only open my mouth to swap feet...

 

[ko99]

Yes, forced air makes heatsinks far more effective. Most of my clients are from the electrical world and forced air is more common than natural convection, but not overwhelmingly so.

Natural convection cooling is pretty straightforward for most 5-10W devices and eliminating fans is great for product reliability and noise. In fact, the advantages of eliminating fans are so attractive, there is a small but growing market for relatively monstrous heatsinks to handle 100W chips (often combined with a heat pipe).

ko (

http://www.ecooling.biz)

 

[snowfire]

Hi, guys! Thanks so much. It's my first time being here. I have tried for a year to join this forum. I have applied more than 4 mail addresses to join it. Haha, so good. I enjoy your post. I have learned something from it. Very good.