Metallurgist? Copper annealing discussion.

[pelle76]

8-ball:

Everything U say makes sense to me. I now remember what I learnd in metalurgics class almost five years ago...

Someone said that the piece would change shape... Nope... As told befor. This happens on a micro-level, not a macro.

8-Ball:

Could u manage to get hold on a fasediagram for Copper??? It would be very nice to see. I think I've got it somewhere but couldent find it.

[BillA]

to the 'no "shape" change' posters:

are you guys talking theory ?
or hands-on experience ?
-> have YOU actually done it ?

I DO speak from experience:
get the bp flat (by measurement) - slowly heat and cool down = NOT flat anymore

I'm sure N8 could add to my comments (or refute them ??)

I understand what is taught (I used to teach several matls courses)

[8-Ball]

I concede.

The likelyhood of the block retaining the exact shape is small, particularly when you are using such accurate equipment.

I wasn't quite sure what angle you were coming from before, whether you were talking about large deviations or not, ie the kind of thing that would warp something beyond use.

8-ball

[8-Ball]

Pelle 76,

Are you after an elemental Cu phase diagram or a Cu-Silver phase diagram.

C110 Copper alloy is almost pure Cu with a bit of silver.

8-ball

[8-Ball]

Cu-Ag Phase diagram (page 2)

Can't find anything for pure copper, though you don't often find phase diagrams for pure elements.

8-ball

[pelle76]

unregistred:

Yepp... Thats true... I thought U talked about big differences.

Question: Would the lapping that would have to be done afterwards cause the copper to go stiffer again??? Not as much as before anywayz...

8-Ball:
Nice, although when looking at the diagram I realised I didn't remeber a shit about those curves... Have to dig my books up after all..

[8-Ball]

Originally posted by pelle76
Question: Would the lapping that would have to be done afterwards cause the copper to go stiffer again??? Not as much as before anywayz...

Probably not a huge amount, but it will have a slight effect. Also remeber that this is the crucial area in the block.

BillA,

How much deformation are we talking about here. Do you have pictures?

8-ball

[8-Ball]

Originally posted by pelle76
Nice, although when looking at the diagram I realised I didn't remeber a shit about those curves... Have to dig my books up after all..

Have you learnt anything about ternary phase diagrams.

They're fun!

8-ball

[pelle76]

Originally posted by 8-Ball
Have you learnt anything about ternary phase diagrams.

They're fun!

8-ball

Yepp... 5 years ago... When I studied for a masters in Mechanics.

We looked at nice laped metall surfaces trough microscpes and compared the grains and other stuff to the TTT diagrams.
The examination of the course was that we got a couple of different metallslabs to look at and then we should write down how they had been treated. I didn't Ace it I tell U but I got through..

[BillA]

nothing I can quantify (before I started spending everything on equip)

bp flatness is terribly sensitive, due also to the presence of 'furniture' on one side only

in any case a probably immeasurable effect from annealing

[Alchemy]

It was my indication in classes that annealing and quenching were not done to pure metals, as the goal was alteration of solid equilibria. I can't think of pure metals with differing crystallization structures that behave as nearly immiscible structural entities, if that's the proper way to describe it.

My text tells me, as I thought, that electron movement is the main executor of thermal transfer.

Alchemy

[8-Ball]

For starters, no-one ever really uses pure metals.

It may be 99.99% pure, but it will still have defects and these defectes will have associated free enrgies and strain fields which will have an effect on electron mobility.

Annealing will provide thermal energy to allow for the diffusion of vacancies and dislocation motion.

This in turn will lead to a coarsening of the microstructure, where by, before it will have been packed full of dislocations and vacancies, and after the smaller number of dislocations will be aranged into low angle grain boundaries, with a much higher proportion of the material having a low defect concentration.

This will improve conductivity.

As a side note, annealing can be and is used to homogenise the composition of materials which have undergone dendritic solidification, or any other solidification mechanism which results in microsegragation, though this is not what we are trying to do here.

8-ball

[gone_fishin]

Photos of crystaline structure

[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.

We come from opposite directions then. I just had an exam that covered how to interpret eutatic phase diagrams, but I know very little about actual material science.

Anyway I agree, defects in diamond or metals will cause a big drop in conductivity, though oddly enough its for completely different reasons.

According to a lecture I had, removing C13 from synthetic diamonds yields a 50% increase in conductivity (diamonds with conductivities beyond 3000w/mk have been made this way IIRC). I wonder how much an effect impurites have on copper . . .

One more thing, whats to stop oxygen from being taken into the crystal when you anneal? I think C101 has to be melted under special atmosphere to prevent this, though I'm only half sure of that.

[8-Ball]

I happened across the answer to this particular problem during some revision I have been doing.

The problem is one of particularly high concentrations of dislocations and point defects (vacancies – lattice sites where an atom is missing) resulting from cold working.

A typical cold worked metal will have a dislocation density of the order of 10^16 per meter squared. This means if you took a plane through a block of copper 1m x 1m, there would be 10^16 dislocations crossing it.

As you can appreciate, this is a stupendously high number. As mentioned before, these defects will affect the electrical conductivity of the material.

So, the solution?

Annealing? Yes, though this is the procedure applied to the material. The actual effects of heating the metal are RECOVERY and RECRYSTALLISATION.

RECRYSTALLISATION is the transformation of a cold worked material (High defect concentrations) to one with essentially no significant concentration of dislocations. The mechanism by which this process works is the passage of high angle grain boundaries.

The driving force for this transformation, (It won’t happen without one), is the reduction of stored energy in the dislocations previously discussed, though this is a very small driving force, relative to the majority of other transformations/reactions.

This transformation needs to be nucleated within the material and then must propagate. Both of these are thermally activated, and provided the sample is heated enough to cause nucleation and propagation, the small driving force of the reduction in stored energy will be sufficient to drive the recrystallisation.

Heating the material to promote recrystallisation will also promote recovery.

RECOVERY is a process which has similar effects, ie the reduction in defect density, though not by the motion of grain boundaries.

There will be mutual annihilation of dislocations above 0.3 of the melting point of the material. This temp is high enough to encourage dislocation glide.

Point defects (vacancies) will also be annihilated by diffusing to edge dislocations. This process results in the climb of dislocations.

The stored energy of dislocations will also be reduced by the arrangement of dislocations into low angle grain boundaries, forming a sub-grain structure.


Ok, that’s the theory of what actually happens. I need not delve further into the processes of nucleation of the recrystallisation, nor the kinetics of the transformation, as for the purpose of this discussion, that which I have described is sufficient.

What does this mean.

I have been able to find experimental data on the recrystallisation of copper samples at various temperatures, shown below, though the exact composition of the Cu is not mentioned. I think it is safe to assume that the lack of any mention of compositions is indicative that this is data for pure copper.

As can be seen, for a recrystallisation temperature of 135.2 degrees C, with 90% recrystallisation being achieved in just under 15 minutes.

This temperature is possible in any oven and can be maintained relatively accurately in a fan oven.

A note on the composition.

The presence of impurities will increase the time required at a given temperature due to solute drag effects slowing down the grain boundary velocity.

So to allow for small quantities of impurities, I would suggest 15 minutes at around 140 degrees C, or 1 hour at 110 degrees C. I don’t know what these temps are in Farrenheit, go find a web converter.

I hope everyone has found this informative and useful.

I apologise for the style of writing. I put this together in word for fear of accidentally hitting the “back” button my mouse, and I only use word for writing essays and pieces of coursework. This is my “please the examiner” style of writing.

8-ball

[EDIT: Image resized]

[Since87]

Excellent writeup.

Thanks 8-Ball.

[pelle76]

Great knowlege 8ball, U da man..

About one of my questions before. The answer struck me after some thought.

"Will the lapping destroy the nice soft copper??"

Answer to myself: Nope, I don't think so. Not if one is beeing gentle about it.

How came this conclusion?:
All the pieces we studied in the microscope at the university were lapped to make a good surface to look at. That didn't destroy those samples so I guess there shouldent be a problem with our lapping either.

[8-Ball]

Originally posted by pelle76
Great knowlege 8ball, U da man..

About one of my questions before. The answer struck me after some thought.

"Will the lapping destroy the nice soft copper??"

Answer to myself: Nope, I don't think so. Not if one is beeing gentle about it.

How came this conclusion?:
All the pieces we studied in the microscope at the university were lapped to make a good surface to look at. That didn't destroy those samples so I guess there shouldent be a problem with our lapping either.

You wouldn't really be able to tell if the lapping had introduced dislocations as they will not be visible under an optical microscope, unless you have specifically etched for dislocations.

The additional energy at dislocations will cause preferential etching at points where dislocations intersect the surface causing etch pits.

If you prepared the samples in the same why I do, you will start with coarse wet and dry paper, roughly 120/200grit, working down to 2000grit, then 6 micron and 1 micron polishes.

There will be a degree of deformation associated with this, (more with the coarse grit wet and dry paper than the polishing), introducing further dislocations, however, this will not be as significant as the dislocation density due to machining or cold working (namely, cold drawn hard copper)

Essentially, you have two options, recrystallise then lapp, risking reintroducing dislocations, or leave it, risking suffering from a "minute" (almost non-existent) deformation in the base.

Either way these will both be minimal effects.

Hope this helps.

8-ball

[Since87]

Originally posted by 8-Ball

Essentially, you have two options, recrystallise then lapp, risking reintroducing dislocations, or leave it, risking suffering from a "minute" (almost non-existent) deformation in the base.

Or...

'Coarse' lap to, say 600 grit.

Recrystallize.

Final lap with 600 grit and higher.

Marking the base before final lapping, might allow a person to crudely determine whether meaningful deformation had occurred.

[Since87]

The thread's subject line has been changed at 8-Ball's request.

Edit: Apparently not.

[8-Ball]

Hasn't changed, but it don't matter.

8-ball

[8-Ball]

Originally posted by redleader
One more thing, whats to stop oxygen from being taken into the crystal when you anneal? I think C101 has to be melted under special atmosphere to prevent this, though I'm only half sure of that.

Redleader,

I'm trying to find the diffusion coefficient for oxygen in copper, then combined with the temperatures and times I recommended above, I should be able to put a diffusion profile together. It's an error function diffusion profile for a fixed surface composition so it shouldn't be that tough to calculate. It would be a nice opportunity to put my diffusion revision to a good use.

That said, note the temperatures I have given are significantly lower than those originally considered, and the penetration depth (stop sniggering at the back) shouldn't be too significant.

Oh, and I'll also need the starting oxygen composition in the electrolytic Cu we use. Anyone?

8-ball

[8-Ball]

Another suggestion,

You could stick your block in boiling water for 2 hours for the same effect.

I'll have a look at the notes I made when I get into the library tomorrow, and check the temperature dependence of diffusion and recrystallisation.

Essentially, to see whether 15 minutes at 140 degrees or 2 hours at 100 degrees will result in more indiffusion of oxygen, though I strongly suspect both processes will have the same temperature dependence.

8-ball

[Since87]

Originally posted by 8-Ball

Oh, and I'll also need the starting oxygen composition in the electrolytic Cu we use. Anyone?

8-ball

I've seen 0.04% oxygen mentioned for C110, but nothing very authoritative. (i.e. worth linking)

[bigben2k]

the chart

Int'l designations