While it’s true conduction and convection are also significant, I don’t agree that 0.7 W is irrelevant. Perhaps radiation isn’t the principle mode but it’s clearly an important factor and becomes more so as the temperature of the flashlight rises. Remember it goes as the fourth power of the absolute temperature.
Thermal diffusivity is a very useful number, however, I would very much disagree that it is the only quantitative figure useful in this discussion. Taking into account only this number, materials such as Pyrolytic graphite, Gold, Silver and Diamond look like excellent candidates.
We have to examine other properties, however. Thermal diffusivity takes into account several factors, but it blurs the absolute value of those factors. For our applications, we don’t want a material that simply moves heat quickly, we also want a material with enough thermal mass to act as a sink when the heat can not possibly be dissipated. Pyrolytic graphite, for instance, has extremely high values of thermal diffusivity, but very low thermal mass. For this reason, if the heat can not be dissipated into the surrounding environment, there is limited thermal mass to act as a reservoir for that heat. In that situation, a material with lower diffusivity but higher mass would actually be a better choice. We also have to consider factors that have nothing to do with thermal transfer. Aluminum, from a purely thermal perspective, appears to be a good choice. It’s downfall, however, is that it’s difficult to solder to. Gold’s downfall is cost. Same with Silver. Diamond is too expensive and too difficult to work with.
Suggesting that the entire discussion of the best heat-sink for an application can be boiled down to one number is akin to suggesting that you can pick the best car by looking at horsepower numbers. If that were the case, we’d all be driving Bugatti Veyron’s.
PPtk
But note that ρ (density) is a factor in thermal diffusivity. I would say that an expression like this in fact makes clear the contributions of the various elements.
Yes, like I said, thermal diffusivity takes into account several factors.
The problem with considering them after they’ve been solved into this neat little number is that
Thermal Conductivity = 5, Density = 4, Heat Capacity = 10
5 / ( 4 * 10 ) = 0.125 Thermal Diffusivity Number
Thermal Conductivity = 1, Density = 4, Heat Capacity = 2
1 / ( 4 * 2 ) = 0.125 Thermal Diffusivity Number
Obviously, these two materials would perform vastly differently. But they have the same Thermal Diffusivity Number.
Great find! I was wondering if anyone had done any experiments. This definitely supports the idea that radiative heat loss is important. I wonder if anyone has made measurements with and without hand-heatsinking (conduction)? I’d guess that helps measurably, too.
"perform" .. Hmm what do you mean by that, got'n example? ;)
if we're looking at the LED bulb temperature as the actual target of investigation than alpha is the single most critical and interesting quantity to look at. but i agree that other practical aspects need to be considered for choosing the optimal heatsink material, e.g. cost of the material.
Wiki has good definition of what a heat sink really is and how its performance can ..etc..blablah.. guys..
radiative loss is quite obviously important. what surprised me is how important the color is!
damn, i love this thread! i am going to try to post up a pic of my wood stove. i have owned and operated a computer business for many years and have recently begun collecting cpu heat sinks. surely intel has done the math and the results are there for everyone to see. my observation is that copper core aluminum heat sinks are the standard for fan based heat sinks while for no fan applications pure copper is utilized and in cramped quarters vented copper with fans are used (notebooks). back to the wood stove, whenever i recycle a pc i now strip out the heat sink and add it to my growing pile on top of my wood stove where a very quiet zalman fan assists the convection currents leeching all the heat i can out of the stove. 
ken
If you get old pennies they are made of copper. So just look for ones before 1981.
actually, 1981 are all copper, too. 1982 pennies could be either. Copper pennies weigh 3.11 grams, whereas the zinc pennies weigh only 2.5 grams.
I have a big bag of 1981 and prior, since I don’t bother weighing 1982 pennies…
Most US pennies before 1982 are bronze, not copper. 95% copper, 5% tin. Enough tin to screw up the thermal conductivity.
would tin really screw it up THAT much?
the problem I have w/ them, is making them flat…
edited:
(original info, had an error http://www.usmint.gov/about_the_mint/fun_facts/?action=fun_facts2)
following is from wikipedia
Years Material
1793–1857 100% copper
1857–1864 88% copper, 12% nickel (also known as NS-12)
1864–1942 bronze (95% copper, 5% tin and zinc)
1943 zinc-coated steel (also known as 1943 steel cent)
1944–1946 brass (95% copper, 5% zinc)
1946–1962 bronze (95% copper, 5% tin and zinc)
1962–1981 brass (95% copper, 5% zinc)
1982 varies, (95% copper, 5% zinc) or (97.5% zinc, 2.5% copper)[6]
1983–present 97.5% zinc, 2.5% copper (core: 99.2% zinc, 0.8% copper; plating: pure copper)[7]
oh my goodness. There things mentioned in this thread that’s way over my head. kreisler H) , math may be simple for you but, it’s like a disease for others.
Another wrench to throw into the discussion: Anodizing! Maybe I read something on another forum over the years but, I’m pretty sure aluminium anodizing has a low thermal conductivity. This is why when maglite heatsinks were first made, they were made a tiney bit bigger than the inner diameter of the tube and the anodizing had to be sanded off for better thermal conductivity… I’m searching forinfo on thermal conductivity of anodizing now.
It really affects the emissivity, too. It looks like anodizing increases it by nearly an order of magnitude:
I have a friend, also a physics teacher, who is building an “off-grid” house and he’s done some experiments with the emissitivities of various materials. He’s found that brushed aluminum is far different than polished aluminum.
for internal parts, where heat is conducted, anodizing could inhibit heat transfer, but imo, minimally
i found the link I posted before when researching whether it is worthwhile to strip external anodizing
it is not, and if that anodizing is black, stripping it will make the host hotter
http://www.molalla.net/members/leeper/coatbar.htm
i tried this myself with a trustfire D1 that had titanium anodizing that I didn’t like, and that is heavily over driven.
I bead blasted the light and it didn’t make much difference. Next I painted it with flat black vht exhaust paint and that definitely helped the light run cooler, for a little while
http://www.dtic.mil/dtic/tr/fulltext/u2/a191755.pdf A lot to digest but interesting!
“The thermal conductivity of typical commercial hard anodic
coatings is about 0.7 V/m/K. The thermal conductivity of thick
anodic coatings produced in an oxalic acid electrolyte is almost
twice (1.3 W/m/iC) that of sulfuric acid-based coatings.”
Yes, apparently, anodizing increases the emissive properties of Aluminum but given its low thermal conductivity, it’s going to stay on the outside but be removed from the inside
yup
sorry I don’t have more numbers from my test on the D1, like the guy in the link did. but really the amount didn’t matter to me, just whether or not being black helped…
Yes, even TINY amounts of impurities in copper can kill its thermal conductivity (like .05% phosphorus can drop it 15%). Many brass/bronze alloys have 1/4 to 1/10 the conductivity of pure copper.
funny how alloys properties are not additive
well, not funny…but interesting.
so, then, texaspyro, what’s your opinion on brass pills vs aluminum?
there are of course many alloys of both, but I’m assuming we’re talking yellow brass and 6061 aluminum…
edit:
ok, probably not yellow brass…
“360 brass” aka “free machining brass”
(60-63% copper, 33.5% zinc, 0.35% iron, 2.4-3.7% lead)
Thermal conductivity: 115 W/m-K@20.0 °C
machinability - 100%
“260 brass” aka “yellow brass” aka “cartridge brass”
(68.6-71.5% copper, 28.5-31.5% zinc, 0.05% iron, 0.07% lead)
Thermal conductivity: 120 W/m-K@20.0 °C
machinability - 30% (based on 360 brass)
“230 brass” or “red brass”
(84-86% copper, 15% zinc, <0.05% lead, <0.05% iron)
Thermal conductivity: 159 W/m-K @20.0 °C
machinability 30% (based on 360 brass)
6061 aluminum
(Aluminum 95.8 - 98.6, Chromium 0.04 - 0.35, Copper 0.15 - 0.40, Iron 0.70, Magnesium 0.8 - 1.2, Manganese 0.15, Silicon 0.4 - 0.8, Zinc 0.25)
Thermal conductivity: 167 W/m-K @77F
machinability 50% (different scale for aluminum alloys)
who wants a red brass flashlight…besides me
an interesting bit I found while searching
“High conductivity* beryllium copper alloys* contain up to 0.7% beryllium, together with some nickel and cobalt. Their thermal conductivity is better than of aluminium, only a bit less than pure copper”
beryllium copper is machinable, but I bet it is to expensive for flashlights…