All about "HEATSINKS" for Laser Builds

[scion]

Ok so, I got a few questions when it comes to heatsinks.

1. What is the best/most effective material to use?
2. Most commonly used material?
3. Alternative materials?
4. How do you determine the size of tge heatsink for your laser?
5. What are some construction methods for making a heatsink?

If you have pictures of your home-made / hand-crafted heatsinks please post them if you can!

I'm trying to find a good method for making heatsinks for some future laser builds!

Thank you in advance for your responses!

 

[Mohrenberg]

1. What is the best/most effective material to use?
-Diamond i believe. If you find a way of making a heatsink of this let me know
2. Most commonly used material?
Aluminum.

3. Alternative materials?
Copper works better than aluminum, but costs more to have made.

4. How do you determine the size of the heatsink for your laser?
this is usually decided by the inside of your host....most heatsinks are made to fill the entire inside of the head of the flashlight.

5. What are some construction methods for making a heatsink?
Mostly machining with a lathe and a mill, this is about the only really effective way to make them, some people make homemade heatsinks from washers, but they don't work nearly as good as a precisely machined heatsink.

 

[roddenberry]

Indeed. Any heatsink made to fit into a flashlight/pointer type has to be milled/bored to the exact size of the module and the "host" to maximize heat dissipation, it would be an incredible stroke of luck to find a metallic part that has these specifications out of the blue... although some copper plumming/brass plumming tubing might come to mind... I'm not sure if copper or aluminium are preferable by weight. Aluminium is much easier to work with, but copper is harder, heavier and looks spiffy

On the other hand, a good labby sink can easily be made out of a CPU radiator with a hole the size of the module drilled into it and with a little fan attached to help cool down the assembly. Really powerful units might require water cooling or these little peltier junction things.

Robert

 

[nikokapo]

1. What is the best/most effective material to use?
-Diamond i believe. If you find a way of making a heatsink of this let me know
2. Most commonly used material?
Aluminum.

3. Alternative materials?
Copper works better than aluminum, but costs more to have made.

4. How do you determine the size of the heatsink for your laser?
this is usually decided by the inside of your host....most heatsinks are made to fill the entire inside of the head of the flashlight.

5. What are some construction methods for making a heatsink?
Mostly machining with a lathe and a mill, this is about the only really effective way to make them, some people make homemade heatsinks from washers, but they don't work nearly as good as a precisely machined heatsink.

EDIT: Deleted incorrect info.

For #4 you can always use some formulas regarding the amount of heat that will be dissipated (energy) and the size of the heatsink (plus shape) needed. The first part will have to be done with a formula I can't remember now, the second one can be easily done with triple integrals to calculate the volume of the heatsink needed (you can always use derivative methods for optimization) if the shape is odd.

 

[scion]

Excellent info! So my best option would be to find some copper tubing that is snug fit all around? Or some type of copper metal that can be cut to spec.

I don't have the tools to bore out some metal. Unfortunately.

Anyone like to post up some pictures regarding different crafted heatsinks?!

 

[Grix]

Everything is correct except #1.

The best material is copper (could be gold or platinum too, but those are better for electrical conductivity). Diamond is PURE carbon, it has 0% metal in it. Carbon is a non-metal and it doesn't conduct heat well (if any at all in this case).

For #4 you can always use some formulas regarding the amount of heat that will be dissipated (energy) and the size of the heatsink (plus shape) needed. The first part will have to be done with a formula I can't remember now, the second one can be easily done with triple integrals to calculate the volume of the heatsink needed (you can always use derivative methods for optimization) if the shape is odd.

Diamond is the best thermal conductor.

Diamond > Silver > Copper > Gold > Aluminium

 

[nikokapo]

Diamond is the best thermal conductor.

Diamond > Silver > Copper > Gold > Aluminium

Edited, I couldn't have been more wrong.

I don't get it, though...aren't electrons not free in diamond?

 

[ZRTMWA]

There is also an copper-diamond alloy metal that was made by scientists. It dissipates heat about 1.5 times better than copper. Pure diamond heatsinks (synthetic or natural) dissipates heat 5 times better than copper.

 

[ColdStl]

For a home made copper style heat sink that will fit the Husky/Pocket Pal type hosts http://laserpointerforums.com/f46/home-depot-sale-cheap-husky-hosts-32256.html Read BobBoyce's post. He "checked" me eliquently when I challenged the size of his copper heat sink vs. my bulkier aluminum one for the same host. Cheap yet effective, and readily available material at HomeDepot. I imagine the same technique could be used for other hosts as well.

 

[bigdalty67]

Edited, I couldn't have been more wrong.

I don't get it, though...aren't electrons not free in diamond?

If i remember correctly from chemistry class the stability of an elements electron configuration does not directly impact its specific heat capacity. What does is the way the atoms are aligned. For instance water has a very high specific heat because of the polarity caused by hydrogen and oxygen molecules. This polarity causes the unique properties of ice and the surface tension of the water. The molecules aren't as free to move as in other compounds or elements which makes it require more joules of energy to raise 1 degree celcius.

Heat has little or nothing to do with the movement of electrons themselves, that impacts electrical conductivity.

Please correct me if I'm way off track! I'd love to learn some more about this stuff.

 

[pullbangdead]

People are confusing some terms with regard to diamond as a heatsink.

A heatsink needs a combination of (1) high thermal conductivity/thermal diffusivity, AND (2) high heat capacity/specific heat. There is no 1 number which will tell you how good of a heatsink it is, you need more than 1.

Diamond is an excellent conductor of heat. Niko, you have to look at how heat is conducted. The most common method of heat transfer is free electrons. In this way, anything metal that has high electrical conductivity will also necessarily have a high thermal conductivity, because those electrons can carry heat as well. BUT, with diamond, there are no free electrons, so no electrical conductivity, but you get thermal conductivity through something called PHONONS. A phonon is basically a vibration, but it's a vibration that can be thought of/approximated as a particle of sorts.

Think of a phonon this way: a solid is like a 3-D network of spheres, all connected to their nearest neighbors with springs (this is actually a pretty good approximation). Say you have a room full of these spheres, all in 3-D, all connected via springs. When you walk up to the edge, and thump one of these spheres, that vibration gets transferred to the other side as a vibration. Thump a sphere on one side, and a vibration will travel through the network, making a sphere on the other side vibrate. That vibration, as it is moving through the lattice, is a phonon.

Now imagine a HUGE set of those spheres, bigger than a football field, still all connected with springs. On one side (the cold side), the spheres aren't moving at all. On the other side (the hot side), you are standing there shaking a sphere back and forth over and over. Now the spheres on the other side are far away, but eventually, they will begin vibrating just like the one you're actually shaking. But not immediately. That's because the phonons, or vibrations, have to transfer the energy to that side through the whole lattice. That's how the heat moves from hot to cold in a diamond, the vibrating atoms transfer energy through the "springy" bonds that hold the lattice together.

Why does it work so well in diamond, and not in other things? First, diamonds have covalent bonds: very strong, very directional. In that case, think of our big network of springs, with very short, tight springs that have a lot of force on them, like you had to stretch them out to twice their length to connect the spheres. But with a metal, you have metallic bonds: much weaker, very non-directional bonds. To approximate these bonds with your ball-and-spring model, it's a lot more like the springs are too long for the distance between atom. The spring is hanging there loosely, or you BARELY had to stretch it at all to connect the atoms. Much less force, a much weaker spring in general. So doesn't it make sense that energy is transferred much more efficiently through the tight springs than it is through the loose springs?

Another factor is how phonons get scattered. Crystals aren't perfect, in a metal there are billions or trillions of dislocations, and grain boundaries, and other things. When a phonon hits these imperfections, if scatters in a different direction, so these imperfections slow down the transfer of heat through phonons. Metals have TONS of these imperfection, while a diamond will have no grain boundaries, and very few dislocations, so the phonons get scattered less. Basically, phonon transport is inefficient in metals, but that lacking is compensated by all those free electrons that can carry heat for you, making them still be good thermal conductors.

If something is a good electrical conductor, it will likely be a good thermal conductor as well, because the charge carriers can carry heat as well. BUT if something is a good thermal conductor, that doesn't necessarily mean it will be a good electrical conductor, because things that are not charge carriers can still carry heat (like phonons).

The rest of the "heatsink equation" is heat capacity/specific heat, but this post answers the questions about diamond as a thermal conductor despite a lack of free electrons.

 

[sbdwag]

Some Nuclear Reactors use liquid sodium metal to transfer heat from the reactor chamber to the steam turbines. So It must have some great heat tranfer or sinking ability but I would'nt worry about using liquid sodium. Its explosive on contact with air and it smells fatal.

regards
sbdwag

 

[nikokapo]

People are confusing some terms with regard to diamond as a heatsink.

A heatsink needs a combination of (1) high thermal conductivity/thermal diffusivity, AND (2) high heat capacity/specific heat. There is no 1 number which will tell you how good of a heatsink it is, you need more than 1.

Diamond is an excellent conductor of heat. Niko, you have to look at how heat is conducted. The most common method of heat transfer is free electrons. In this way, anything metal that has high electrical conductivity will also necessarily have a high thermal conductivity, because those electrons can carry heat as well. BUT, with diamond, there are no free electrons, so no electrical conductivity, but you get thermal conductivity through something called PHONONS. A phonon is basically a vibration, but it's a vibration that can be thought of/approximated as a particle of sorts.

Think of a phonon this way: a solid is like a 3-D network of spheres, all connected to their nearest neighbors with springs (this is actually a pretty good approximation). Say you have a room full of these spheres, all in 3-D, all connected via springs. When you walk up to the edge, and thump one of these spheres, that vibration gets transferred to the other side as a vibration. Thump a sphere on one side, and a vibration will travel through the network, making a sphere on the other side vibrate. That vibration, as it is moving through the lattice, is a phonon.

Now imagine a HUGE set of those spheres, bigger than a football field, still all connected with springs. On one side (the cold side), the spheres aren't moving at all. On the other side (the hot side), you are standing there shaking a sphere back and forth over and over. Now the spheres on the other side are far away, but eventually, they will begin vibrating just like the one you're actually shaking. But not immediately. That's because the phonons, or vibrations, have to transfer the energy to that side through the whole lattice. That's how the heat moves from hot to cold in a diamond, the vibrating atoms transfer energy through the "springy" bonds that hold the lattice together.

Why does it work so well in diamond, and not in other things? First, diamonds have covalent bonds: very strong, very directional. In that case, think of our big network of springs, with very short, tight springs that have a lot of force on them, like you had to stretch them out to twice their length to connect the spheres. But with a metal, you have metallic bonds: much weaker, very non-directional bonds. To approximate these bonds with your ball-and-spring model, it's a lot more like the springs are too long for the distance between atom. The spring is hanging there loosely, or you BARELY had to stretch it at all to connect the atoms. Much less force, a much weaker spring in general. So doesn't it make sense that energy is transferred much more efficiently through the tight springs than it is through the loose springs?

Another factor is how phonons get scattered. Crystals aren't perfect, in a metal there are billions or trillions of dislocations, and grain boundaries, and other things. When a phonon hits these imperfections, if scatters in a different direction, so these imperfections slow down the transfer of heat through phonons. Metals have TONS of these imperfection, while a diamond will have no grain boundaries, and very few dislocations, so the phonons get scattered less. Basically, phonon transport is inefficient in metals, but that lacking is compensated by all those free electrons that can carry heat for you, making them still be good thermal conductors.

If something is a good electrical conductor, it will likely be a good thermal conductor as well, because the charge carriers can carry heat as well. BUT if something is a good thermal conductor, that doesn't necessarily mean it will be a good electrical conductor, because things that are not charge carriers can still carry heat (like phonons).

The rest of the "heatsink equation" is heat capacity/specific heat, but this post answers the questions about diamond as a thermal conductor despite a lack of free electrons.

I have to thank you for this AWESOME explanation you gave me

Thanks man! I can't believe I just came from taking a "Material Science" exam 5 hours ago at college and I didn't know this (though we've only studied metals so far).
Excellent analogies

However, I don't understand how heat is generated by phonons. Are the vibrations transformed into heat because of the vibration of ions? (are diamonds ionized?). I mean, how is heat transferred from one end to the other if electrons aren't free, do ions crash with each other and transfer energy?

 

[Flaminpyro]

@ SCION heres some different types of heatsinks I recently made.

 

[pullbangdead]

Some Nuclear Reactors use liquid sodium metal to transfer heat from the reactor chamber to the steam turbines. So It must have some great heat tranfer or sinking ability but I would'nt worry about using liquid sodium. Its explosive on contact with air and it smells fatal.

regards
sbdwag

I don't know this for certain, but I've been told that liquid sodium is actually the best elemental electrical conductor there is, higher than silver, gold, any of those elements. That electrical conductivity would give it a very high thermal conductivity, as well.

The other advantage of a liquid though, is of course the ability to circulate it, so you get convection along with conduction for heat transfer.

I have to thank you for this AWESOME explanation you gave me

Thanks man! I can't believe I just came from taking a "Material Science" exam 5 hours ago at college and I didn't know this (though we've only studied metals so far).
Excellent analogies

However, I don't understand how heat is generated by phonons. Are the vibrations transformed into heat because of the vibration of ions? (are diamonds ionized?). I mean, how is heat transferred from one end to the other if electrons aren't free, do ions crash with each other and transfer energy?

Heat is not generated by phonons, heat is transferred by phonons. The heat is already there on the hot end. The simplest case is a rod that has a hot end and a cold end (let's say a rod of something like diamond, no free electrons, but good phonon transport). On the hot end, the atoms have higher average kinetic energy. That kinetic energy is in the form of a vibration, each atom is vibrating with some energy, and the average vibrational energy is the temperature. For reference and to give you the scale here, an average energy of 1 eV gives you a temperature of ~11,605K. So yeah, 1 atom at 100C/373K has a VERY low vibrational energy, and this makes perfect sense, since you have so many atoms. It's not a big vibration, it's tiny.

The higher the "temperature" (in quotations because I like to think of temperature as an average, and an average has little meaning when talking about one atom) of an individual atom, the more it's vibrating. The fact that the atom is bonded to another atom means the bonded atom will ALSO start vibrating, due to the first one vibrating. If one atom moves left, the one next to it is pulled left. The first one moves back right, the other is pulled back right. The bonds we're talking about are pretty much always some form of a Coulombic attraction, so it's the electrical attraction/repulsion making each atom affect its neighbors as it displaces. an atom displaces, which displaces its neighbors because atoms always want to be at some set equilibrium distance from one-another. That is carried out over and over, as that vibration is transferred to every adjacent atom all the way down the line. The energy of the hot side is in the form of the atom vibrations on that side, and through phonons, you end up with atoms vibrating on both sides, the hot and cold side equalize to the same temperature, as those propagating vibrations transfer that kinetic energy from high-energy atoms to low-energy atoms.

Atoms don't necessarily "crash" into one another, but every atom it always vibrating around some average position. Since each atom has a bond with its nearest neighbors, one atom vibrating will affect all the adjacent atoms, over and over again through the whole crystal. This process is very efficient in diamond because the bonds themselves are so strong: instead of a loose spring like in a metal, the spring between carbon atoms in diamond would be a very stiff spring, almost a solid bar of a connection. Think about a very loose spring between 2 balls: moving one ball doesn't move the other very much, because the spring is so "un-stiff". But you put a very stiff spring, or even a stick, between the two, and you get one atom being very heavily influenced by the vibration of the other. So diamond transports its phonons much more efficiently, with its stiffer bonds.

The energy "comes from" whatever your heat source is, but then is transferred from hot to cold through phonons or electrons. Well, those are the main 2 transfer mechanisms in solids. You can also sometimes get ions to carry heat if you have free ions in the solid, or you can even transfer heat within a solid through radiation under the right circumstances.

As far as if the carbon atoms in diamond are "ionized", that's not really what is in play with phonon transport, phonon transport is the physical interaction of neighboring bodies that interact Coulombically as they vibrate. Are the atoms ionized? Not really, the carbon atoms are in hybrid orbitals (sp3 I think?), so each atom is not technically in the "ground state" that would exist if a carbon atom were isolated all by itself, but I think of ionized as having an electron removed entirely from the atom such that it is no longer localized, which doesn't happen in pure diamond. All the atoms are still localized to their bonds in pure diamond/carbon, and the solid as a whole is in a ground state, seeing as how it is extremely stable and definitely in a low-energy state.

 

[nikokapo]

I don't know this for certain, but I've been told that liquid sodium is actually the best elemental electrical conductor there is, higher than silver, gold, any of those elements. That electrical conductivity would give it a very high thermal conductivity, as well.

The other advantage of a liquid though, is of course the ability to circulate it, so you get convection along with conduction for heat transfer.



Heat is not generated by phonons, heat is transferred by phonons. The heat is already there on the hot end. The simplest case is a rod that has a hot end and a cold end (let's say a rod of something like diamond, no free electrons, but good phonon transport). On the hot end, the atoms have higher average kinetic energy. That kinetic energy is in the form of a vibration, each atom is vibrating with some energy, and the average vibrational energy is the temperature. For reference and to give you the scale here, an average energy of 1 eV gives you a temperature of ~11,605K. So yeah, 1 atom at 100C/373K has a VERY low vibrational energy, and this makes perfect sense, since you have so many atoms. It's not a big vibration, it's tiny.

The higher the "temperature" (in quotations because I like to think of temperature as an average, and an average has little meaning when talking about one atom) of an individual atom, the more it's vibrating. The fact that the atom is bonded to another atom means the bonded atom will ALSO start vibrating, due to the first one vibrating. If one atom moves left, the one next to it is pulled left. The first one moves back right, the other is pulled back right. The bonds we're talking about are pretty much always some form of a Coulombic attraction, so it's the electrical attraction/repulsion making each atom affect its neighbors as it displaces. an atom displaces, which displaces its neighbors because atoms always want to be at some set equilibrium distance from one-another. That is carried out over and over, as that vibration is transferred to every adjacent atom all the way down the line. The energy of the hot side is in the form of the atom vibrations on that side, and through phonons, you end up with atoms vibrating on both sides, the hot and cold side equalize to the same temperature, as those propagating vibrations transfer that kinetic energy from high-energy atoms to low-energy atoms.

Atoms don't necessarily "crash" into one another, but every atom it always vibrating around some average position. Since each atom has a bond with its nearest neighbors, one atom vibrating will affect all the adjacent atoms, over and over again through the whole crystal. This process is very efficient in diamond because the bonds themselves are so strong: instead of a loose spring like in a metal, the spring between carbon atoms in diamond would be a very stiff spring, almost a solid bar of a connection. Think about a very loose spring between 2 balls: moving one ball doesn't move the other very much, because the spring is so "un-stiff". But you put a very stiff spring, or even a stick, between the two, and you get one atom being very heavily influenced by the vibration of the other. So diamond transports its phonons much more efficiently, with its stiffer bonds.

The energy "comes from" whatever your heat source is, but then is transferred from hot to cold through phonons or electrons. Well, those are the main 2 transfer mechanisms in solids. You can also sometimes get ions to carry heat if you have free ions in the solid, or you can even transfer heat within a solid through radiation under the right circumstances.

As far as if the carbon atoms in diamond are "ionized", that's not really what is in play with phonon transport, phonon transport is the physical interaction of neighboring bodies that interact Coulombically as they vibrate. Are the atoms ionized? Not really, the carbon atoms are in hybrid orbitals (sp3 I think?), so each atom is not technically in the "ground state" that would exist if a carbon atom were isolated all by itself, but I think of ionized as having an electron removed entirely from the atom such that it is no longer localized, which doesn't happen in pure diamond. All the atoms are still localized to their bonds in pure diamond/carbon, and the solid as a whole is in a ground state, seeing as how it is extremely stable and definitely in a low-energy state.

A big thanks, man. That cleared everything out for me