Annealing of a metal is a process whereby the metal is subjected to high temperatures for a period of time, and then allowed to cool, either gradually, or quickly, depending on exactly the type of metal and what you wish to do with it.Generally speaking, this is done to make a metal softer, more ductile, and / or to reduce any internal stresses that may have occurred as a result of forging the metal, or any strain hardening processes applied to it. As you may or may not know, metals are composites of a bunch of tiny little crystals stuck together. Now, the way it works these crystals are like balls stacked together in a pile. Depending on just how they're stacked, there will be one or more planes along which rows of the balls could slide relative to each other. For a nice example of how this effect works, see this page. Now, the boundaries between the crystals, this is where the crystal will stop sliding in the same direction. Thus, the more boundaries, the greater resistance to movement the metal will have. I.E., it'll be harder. The smaller the average crystal grain size, the more boundaries, the stronger the metal is, and conversely, the less ductile it will be.

This is where annealing comes in. In general, when we heat metals hot enough, the crystal grains will start growing together. This takes some time, anywhere from 10 minutes to 3 hours, depending on the material, the shape, and which specific annealing process you wish to perform. Its a bit complicated for some metals, but this is pretty much the entire annealing process for a lot of metals that have simple structures. Much like your regular solids / liquids / gasses, when we're talking about formations of metal, we refer to specific formations of metal as phases. Phases for alloys are generally tracked on a temperature / concentration (of one of the components of the alloy) graph. This doesn't actually track what phases the substance will be at that particular temperature and concentration, its the phase which will be stable. Depending on exactly what you do to the metal, what you get and what is stable can be very different, because the rate at which it changes back is extremely slow. A quick non metallic example that everyone can relate to is a diamond. A diamond is not forever, because it is not the phase of carbon that is stable at room temperature and atmospheric pressure. That would be graphite. But because the process requires a great deal of activation energy, it is very slow, thus your engagement ring isn't going to be suitable for writing with for quite some time.

This is the point of allowing the metal to cool slowly. What happens sometimes when you quickly cool a metal is that it gets stuck into a higher energy state. This is why when a blacksmith is making a sword, he heats it nice and hot, and then quenches the sword in water, very quickly cooling it down, and trapping the phase of the steel into martensite, which is very hard, and also brittle. When we cool it slowly, however, it will not be trapped in any formation other than the formation indicated upon the phase diagram. That is the general end result of an annealing process, a metal with larger grain sizes, of the phase that it should be, without any internal stresses between crystals caused by deformations. At this point I will note that one of the side effects of annealing is to undo the effects of strain hardening, negating the increase in strength and the decrease in ductility caused by that process.

Some materials are simpler than this. No matter how fast you cool them, they will form the same phase. In this case, the purpose of annealing is not to change the phase at all, but simply to adjust the grain size, and reduce any internal stresses. Some examples of these blessedly easy to anneal materials include brass, copper, silver, zinc, and a number of other non-ferrous metals.


Specific Annealing Processes

Stress Relieving

Although sometimes not considered a true annealing process, it is quite similar, so I'll group it here anyways. Basically, you heat up the metal. To what temperature depends on exactly what metal we're talking about, it ranges (From the figures I have available, which isn't much at the time since I don't have my textbooks with me.) from 150 °C for Magnesium, to 1215 °C for certain alloys of Nickel, with steel, one of our favorite substances, falling in the middle at between 500–625 °C. This temperature is held for a while, usually around an hour. After that is is slowly cooled. One classic way blacksmiths accomplished this was to plunge a piece into a bucket on sand or lime, which would stick to the piece being worked, and insulate it as it cooled.

Now, the temperature doesn't reach as high as it does in most other annealing processes. This means that the changes to the microstructure are not nearly as severe, and it retains most of the properties remain the same. The strength will drop a bit, and the ductility will rise a bit, but for the most part the main difference is in the reduction of internal stresses. These happen during strain hardening, or when the part in question is being formed, via forging, machining, or whatever it is that they made it with. From what I understand, they're basically little parts of the metal that have been dislocated in such a way that they're pushing, or pulling against the rest of the structure. Needless to say, when you want to apply a load to the part made out of this metal, these spots are the parts that are most likely to give way. We want to reduce the chances of this happening. This is why we go through this rather expensive process with a number of important mechanical parts.

Full Anneal

For a full anneal, you heat the piece to a higher temperature. In steel, that ranges from 950 °C–760 °C, or 815 °C–1215 °C for various alloys of Nickel. Once heated up to that temperature, you allow the metal to "soak", until it has completely converted into the phase which is expected at that temperature on the phase diagram. In all cases, this will be a phase with a single structure type, in steel's case a phase called Austenite. Much like sugar dissolving in tea when its warm, when metals warm up, any foreign particles are "dissolved" into a single type of phase. It should be noted that this happens a much lower temperatures than the melting point of the alloy.

After soaking for long enough, the temperature is very slowly dropped. This may require a rather expensive furnace, as its a bit more complicated than a simple on-off switch. You drop the temperature about 20 °C per hour, until it has been dropped past a certain temperature (for steel: 727 °C), after which you can simply air cool the component.

There is another very closely related process, known as Normalizing. The only difference between the two is that after soaking, instead of a controlled cooling, the parts are allowed to air cool the entire way. This results in a finer structure (for steel: Fine Pearlite, instead of Coarse Pearlite), which will be stronger and less ductile. As well, the fully annealed piece will have a completely homogeneous structure throughout, as the entire piece cools at the same rate. On the other hand, the normalized piece will cool at different rates throughout, depending upon thickness, geometry, and whatever types of convection currents arise to cool the metal.

Spheroidization

This is an annealing process exclusive to high carbon steels. To start off, you take steel in a pearlite structure, which would be pretty much any normal steel. You heat it up to between 700 °C–800 °C, and hold it there for several hours. What will happen is that the carbon rich Cementite will begin to separate from the Ferrite structure, clumping together and forming small blobs, instead of the alternating layer pattern of the pearlite. Cementite is harder than Ferrite, so what you'll have is dark gobs of hardness in a softer, shiny matrix.

Now, left on its own, this arrangement is less useful than a straight pearlite mix. When it does become useful, however, is if you are planning on manufacturing something through a process which squeezes or stretches the metal, such as extrusion, or forging. When this happens, the globules of cementite get stretched, and elongated. These tiny hard globs become thin hard spikes stuck in the part. This increases the strength of the part, and if you form the part the right way, you can arrange it so that it is especially strong along the axis which the part will be subjected to stress. This process is often used for tools, such as screwdrivers and wrenches.

Torch Annealing

And finally, annealing on a much smaller scale than I've been talking about. The idea is you take some kind of portable heat source, such as an oxyacetylene torch, and use it to heat up the metal. Then, you allow it to cool. Naturally, this process is a lot less exact than one done by a temperature controlled furnace, and the results will not be as good, but it is still used by a wide range of people, from welders on the job, to home ammunition manufacturing enthusiasts. It doesn't require nearly as much infrastructure, nor does it require you to transport the piece to be annealed to wherever your furnace is. However, the results will not be quite as good.


Naturally, this is a rather simplified overview of a rather complex topic. There are many other factors to consider, such as combating surface oxidation which occurs at the temperatures required to anneal many Nickel alloys, or some rather complicated annealing processes involving Titanium, where they attempt to include creep, in order to form the metal. But, that's a bit beyond the scope of this node. To recap, annealing is when we heat a metal for a while, in order to make it more ductile, and less likely to break on us. After all, we generally prefer our bridges to stay standing.

smartalix says Here's a novel application of annealing: mbridgetech.com
smartalix says A one-sentence summary: This company makes a programmable resistor by controlled annealing of a microscopic ingot on a silicon wafer to change its conductivity characteristics.


Sources:
rep_movsd. "Heat treatment of steel," Everything2.com <everything2.com/index.pl?node_id=1537663> (September 27, 2005).
wrinkly. "martensite," Everything2.com <everything2.com/index.pl?node_id=1281338> (September 27, 2005).
Gregory Clarke. "Pearlite," Everything2.com <everything2.com/index.pl?node_id=296665> (September 27, 2005).
Gregory Clarke. "austenite," Everything2.com <everything2.com/index.pl?node_id=1317900> (September 27, 2005).
Gregory Clarke. "Strain hardening ," Everything2.com <everything2.com/index.pl?node_id=1369443> (September 27, 2005).
INI International. "Welcome to the World's Most Comprehensive Steel Properties Database," Key to Steel. 2005. <www.key-to-steel.com> (September 27, 2005).
INI International. "Welcome to the world's most comprehensive METALS DATABASE," KEY to METALS. 2005. <www.key-to-metals.com> (September 27, 2005).
Blackstone Mill. "Blacksmith Tools," Blackstone Mill ROCKS & FOSSILS TOOLS & ANTIQUES. <www.blackstonemill.com/Blktools.htm> (September 27, 2005).
eFunda, Inc. "Heat Treatments: Softening: Annealing," eFunda: The Ultimate Online Reference for Engineers. 2005. <www.efunda.com/processes/heat_treat/softening/annealing.cfm> (September 27, 2005).