Metallurgist? Copper annealing discussion.

[Since87]

Some of us have been discussing copper annealing in this thread.

According to this list of thermal conductivities hard drawn copper (which is preferable to use for machining) has only 89.5% of the thermal conductivity of annealed copper.

Anyone know enough about this stuff to suggest a good home annealing technique?

[Kwissus]

Im not claiming to be a pro in any way, but an Oxy-acetylen torch would do the trick, If you use less oxygen than you normaly do. (This, of course, to maximise the amount of "airborn" oxygen to burn, lessening(is that a word?) the amount of oxide on the copper)

[8-Ball]

Considering I'm doing a degree in metallurgy and the science of materials, I should be able to help out.

It'll probably take me a few days though, as I don't have all of my notes with me and tbh, my finals are at the top of my list of priorities.

I'll try and get back soon.

8-ball

[8-Ball]

Originally posted by Since87
According to this list of thermal conductivities

Errm, that list is is for electrical conductivities, and although they are often related, that is not always the case. Again, I'll have to check some books and notes when I get home.

8-ball

[Since87]

Originally posted by 8-Ball
Errm, that list is is for electrical conductivities,

You're right. Sloppy on my part. I'd googled for a list of thermal conductivities, saw "Relative Conductivity" at the top of the column and didn't look further.

Still, interested in what you can find out 8-Ball.

[jaydee]

I went to my engineer buddy's house a couple weekends ago and he brought up that annealed copper is not only easier to mill but has a better conductivity and it might be worth looking into. Next time I see him I will ask him about the process. He does this shit for living. He was designing a 75ft long heat pipe when I was there. You should see all the cool stuff in their dumpster. Miles of coils and heat exhanger stuff. What pisses me off is the guy builds this stuff for a living yet he bought a Dell. :shrug:

[golovko]

Originally posted by Since87

According to this list of thermal conductivities hard drawn copper (which is preferable to use for machining) has only 89.5% of the thermal conductivity of annealed copper.

Anyone know enough about this stuff to suggest a good home annealing technique?

An annealed material has better thermal conductivity than that of a cold worked material because it has larger grains which transmit heat better than the smaller grains found in a cold worked (in this case hard drawn) material.

The are many variables that affect what temperature a material should be annealed at, but it is typically between 1/3 and 1/2 of its melting temp (pure copper's Tm is 1083 C). So to anneal a block of copper (assuming its relatively close to pure) youre going to have to heat it in the range from 360 C to 540 C. A torch should work assuming it burns that hot (I'm not too familiar with torches or welding). I'll speak with one of my professors today who may be able to recommend a better way to "home-anneal" a block of copper.

[Since87]

Hmm.

My impression was that machinists preferred hard drawn copper because it wouldn't 'gum up' as much when milled.

Annealed is much easier to bend into shape though.

[pelle76]

Golovko:

Yep.. that sounds right. I remember talking about 400 degrees when we looket at copper in metalurgy class (only basics)

Ok... So U say that thermal conductivity increases with the grainsize. Nice to have that settled out.

It will work to heat copper with a acetylene/oxygen torch. I've done it. The copperplate went from rock hard (some spooky alloy) to soft as a kitten .

If I remeber correct... As long as U heat the copper more than say the 400 degrees then there is no upper roof for the temperature. As long as U let it cool at a slow rate U will get big corns???
What I'm trying to say is that one dont have to be so precasious about heating it to much???

[8-Ball]

Up to around 600 degrees is pretty good

[jaydee]

Originally posted by Since87
Hmm.

My impression was that machinists preferred hard drawn copper because it wouldn't 'gum up' as much when milled.

Annealed is much easier to bend into shape though.

Bending into shapes is what brought up out conversation on it. He said they use annealed sheet metal in some things because it is a lot easier to bend and some other stuff I already forgot. He did say it should mill easier though. Maybe those larger grains golovko mentioned makes it easier to mill. It is the density that really makes it hard to mill and not so much the softness. If for some reason the annealing makes it softer is should allow the cutter to remove more material faster even at the same density. or maybe i just showed how ignorant I am on the subject.

[LiquidRulez]

Well, If it is softer after annealing...Then wouldnt it bend around the die ,when mounted on a CPU?Or easily gain a impression of the much harder surface of the die ,on the WB base??
Furthermore.........wouldnt the base of the block itself bend when trying to achieve the proper amount of mounting force on the die???

It seems this is something that would not be good for what we are using it for.

[Since87]

Originally posted by LiquidRulez
Well, If it is softer after annealing...Then wouldnt it bend around the die ,when mounted on a CPU?Or easily gain a impression of the much harder surface of the die ,on the WB base??
Furthermore.........wouldnt the base of the block itself bend when trying to achieve the proper amount of mounting force on the die???

It seems this is something that would not be good for what we are using it for.

Well even 'soft' copper is pretty stiff compared to the fiberglass of the MOBO. I don't think there is much reason for concern, unless the block baseplate is very thin, with no mechanical support above the die.

[myv65]

Originally posted by jaydee116
or maybe i just showed how ignorant I am on the subject.

LOLOL. Grain size, hardness, and yield strength in metals are intimately related topics. Density of copper (or other metals) is virtually unaffected by these items and does not have a direct influence on machineability.

Annealing is only one half of the equation, too. The other half is controlled cooling, though I don't have any feel for its impact with copper. In alloyed steels and iron, the rate of cooling can affect the structure and material properties dramatically. As an example, the paper industry uses "chilled iron" rolls. The roll shells are formed in a single pour, but the exterior is rapidly cooled. While chemical composition is the same throughout, the outer ~3/4" is white iron (very hard, low tensile strength, low thermal expansion, extremely difficult to cut) the remainder is gray iron (soft, decent tensile strength, good conductivity, relatively ductile, easy to cut). The exterior resists wear while the interior gives strength and thermal conductivity. The impact of carbon on steel is similar in that higher carbon content allows much higher hardnesses to be reached, but only with the right cooling rate. Hence the variations of air cooled vs water cooled vs brine etc. Just to make certain laypersons don't figure any of this out, metal workers and engineers call it "annealing and quenching" rather than heating and cooling.

Specifically with regard to thermal conductivity and grain size, I'd point out that the very act of making a block destroys some of the effort. Any machining operation done at "low" temperature results in cold-working. This is basically causing the metal to yield, which severs existing grain boundaries leaving smaller grains behind. This is why a soft metal is often harder after it's been machined. It's also why the hardness of a wire increases once you've bent it back and forth a few times (and why it will eventually fail in a brittle nature, even though the raw material may have been ductile).

If you were talking about steel, I could point to some TTT diagrams of relevance. As it is, I'll shut up now and leave it to those with better information at hand, like 8-ball.

[pelle76]

The TTT diagram (fasediagram) of copper is not as complex as for Steel. Steel is an alloy or composite (carbon, nickel, crome etc...)
That makes it complex and the heating and cooling must be quite exact to get the steel ure after.

One could easily soften copper after the milling by heating it up. The question is: Would it serve or worsen the chilling of the die???

Here are two aspects to consider (I bet there is more)
1. Thermal conductivity
2. Softer materials makes better contact with core.

[hydrogen18]

id like to say that copper cooling rates while annealing makes a difference. The quicker the cooling causes the copper oxide to break off quickly leaving behind the layer of cupric oxide(which is photovoltaic i might add). Then just remove the cupric oxide as mentinoed b4. Hmm, gives me an idea...i might try cooling it with liquid nitro.

[hara]

Originally posted by hydrogen18
id like to say that copper cooling rates while annealing makes a difference. The quicker the cooling causes the copper oxide to break off quickly leaving behind the layer of cupric oxide(which is photovoltaic i might add). Then just remove the cupric oxide as mentinoed b4. Hmm, gives me an idea...i might try cooling it with liquid nitro.

What does this mean for us. I don't understand

[bigben2k]

Originally posted by hara
What does this mean for us. I don't understand

Nothing, unless you work in a foundry, or you're making safety glass.

[8-Ball]

myv65, your description of how work hardening takes place is a little off, though the result as you have described is the same, and the way it works is less important.

OK, I'm gonna run some stuff off the top of my head about how and why annealed copper might have a "slightly" higher thermal conductivity. (Please note that we don't really get taught about heat flow or thermal properties, this is just putting ideas together.) Also note, that if none of this makes sense, then I probably won't ave the time to take it any further.

Heat flow through most close packed solids, ie metals, is by phonons, lattice vibrations, which travel through the material. these vibrations are going to be largely affected by the bond strength/length between atoms.

Now this is where the hard/soft comes in.

No material is perfect!

The microstructure is not continuous. The copper we handle is polycrystalline, ie the material is split into grains, regions of crystalline material surrounded by other grains of material with a different orientation.

If that weren't bad enough, there are defects within the microstructure, whether they be missing atoms (vacancies), impurities, or dislocations. These are the main three we need to worry about.

Firstly, dislocations. Believe it or not, it is these defects which give high strength steels their strength. Drawn eutectoid steels have been produced specifically to have the highest dislocation density possible. These make up the cables on suspension bridges, tyre chord and so on.

When a material is subjected to a stress, a number of things begin to happen. Firstly elastic deformation. This is simply stretching of the bonds and is not permanent.

Once the elastic limit is reached, (yield strength), plastic deformation begins. This is when things get messed up.

A dislocation is a line defect. Imagine a stack of crystal planes, atoms packed together as tightly as possible. Now imagine half a plane wedged in there. The line at the edge of this half plane is a dislocation. Materials deform largely by various mechanisms of dislocation motion. Occasionally, dislocations can pass over one another and get pinned. Other dislocation formations can continuously generate new dislocations.

To cut a long story short, through deformation, a mass of dislocations are formed which tangle up with one another, grain boundaries, impurities, or anything else which will get in the way.

This is the process of work hardening.

Machining copper is essentially work hardening, and hard drawn copper, as discussed here has been deformed hugely at room temperatures to refine the microstructure.

All of these defects, grain boundaries included, have a stress field associated with them, distortions in the lattice surrounding them as the lattice attempts to accomodate this defect.

For a phonon trying to traverse a material, high defect concentrations is a nightmare.

So we need to aneal.

Dislocation/vacancy motion, essentially diffusion, are thermally activated. By heating the material dislocations move around, alingning themselves into "sheets" of dislocations forming low angle or tilt grain boundaries within grains. Vacancies diffuse to and from dislocations as they climb, and with time, the microstructure will coarsen, becoming large and fairly uniform.

A material in this state will allow the phonons to travel much more easily, thus increasing the thermal conductivity, at the expense of strength, hardness and so on.

It is for this reason that we have these different types of Cu, depending on the application requirements.

As a side note, diamond is such a good conductor because of it's near perfect crystalline structure high stiffness covalent bonds.

If this has gone over your head, please don't expect me to explain the whole thing in great detail.

To those who may have understood most of it, if there are small points which don't make sense, ask away, though I warn you, I will be busy, tired and a little grumpy over the next few weeks.

Again, let me emphasise that this is my attempt at putting together a reasonable explanation of what is happening, based on what I have learned. It was intended as an informative post aimed at those who want to learn more.

As for the annealing itself, I don't know off the top of my head at what temp the Cu should be annealed, or for how long, or indeed how it should be cooled. I will say that for near pure materials, the cooling rate is less of a critical factor as it is with high alloy steels.

Hope folks have benefitted from this.

8-ball

[bigben2k]

Nice. That actually made sense to me, but you clearly have a better understanding of it than I do.

Did you notice my safety glass line (above)?

[Since87]

Yes, thanks 8-Ball.

[golovko]

Originally posted by 8-Ball
Considering I'm doing a degree in metallurgy and the science of materials...

We may be majoring in roughly the same thing - my major is Materials Science & Engineering.

[redleader]

Heat flow through most close packed solids, ie metals, is by phonons, lattice vibrations, which travel through the material. these vibrations are going to be largely affected by the bond strength/length between atoms.

Metals are not diamonds. Conduction in metals is due to electrons passing between atoms through the conduction band. Phonons are in strongly bonded covalent crystals like diamond. That is why thermal and electrical conductivity are so closely related in metals, but not in diamond.

Allow me to quote my text book:

However, it is a common observation that the conductivity of a metal decreases as the temperature increases. The mobility in a metal is limited by electron scattering. This is dominated by impurities and crystal defects at low temperatures, but is caused mainly by lattice vibrations (phonons) above room temperature. Since thermal vibrations increase with temperature, the electronic mobility and, hence, overall conductivity decreases as the temperature increases in a metal. Conductivity in a metal is also very sensitive to impurities and other defects which scatter the electrons. The non-directional metallic bond allows defects to be accommodated with relative small increases in energy.

and again from another chapter:

Metals also have very high thermal conductivity, but this is due to the heat carried by the electron gas.

[BillA]

before you guys get too carried away,
remember that annealing is also called 'stress relieving'
trust me, the piece WILL change shape

[8-Ball]

redleader,

As I said, I've not been taught about thermal conduction, though the principles I have outlined will have the same effect on electron mobility. Essentially, there are two extremes.

No defects, good conductivity, crap strength.

Lots of defects, crap conductivity, great strength.

You would not believe how low the dislocation densities are in modern semcionductor wafers.

BillA

annealing can change the workpieces shape, though it doesn't necessarily have to.

Granted some work pieces may have large scale stresses, but the stresses implied with stress relieving are microscopic, as opposed to macroscopic.

The point of annealing is to give diffusion (thermally activated) time to lower the free energy of the system.

Vacancies, defects, dislocations and impurities have free energies associated with them due to their surrounding strees fields. By annealing, the material is rearranged into sub grains with few defects. A large number of dislocations will be anhiliated by collision with a dislocation of the same type (edge or screw) and an opposite burger's vector. (A vector quantity which defines the effect of the dislcocation). A large number of vacancies will diffuse to grain boundaries where they will naturally segregate into locations where atoms would experience high stress fields.

Essentially, annealing is the thermal diffusion of matter/dislocations/vacancies, so as to lower the overall energy.

This allows the material to maintain the current defined shape without the high energy concentrations associated with dislocation locks, twins and so on.

If you like, I will check this with one of my lecturers.

8-ball