Thanks for the info Mike. I have a precision furnace so I can do this. I assume I do it on a A rivet as opposed to an AD rivet?
Please don't use 'A' rivets.
The point here is to soften an 'AD' rivet so it is easier to squeeze, in a way that it will recover its strength.
Age-hardening (also called precipitation hardening) of aluminum takes place as an alloying agent, copper in the case of 2000-series aluminum, precipitates out of the aluminum crystal solution. The random dispersal of copper atoms in the interstitial spaces in the crystal interfere with the movement of aluminum atoms as the crystal is strained. (actually it is minor defects in the crystal called dislocations that move, by making individual atoms adjust their position around the dislocation). By interfering with the movement of dislocations, it takes more force to move them. Thus the crystal is stronger.
When you solution-treat the aluminum, what you are doing is putting the copper atoms back into solution in the crystal. They take the place of aluminum atoms in the matrix and do not interfere with the movement of dislocations. So the material acts 'soft'.
Over time, the copper atoms precipitate out of the crystal again, regaining strength.
This process is quite different from annealing. Annealing is done at a temperature BELOW the solution treating temperature. Typically 750F for aluminum. It does not dissolve the copper atoms back into solution. What it does is allow the aluminum crystals to 'relax'. Internal stresses are released by aluminum atoms moving or adjusting their lattice linkages. Dislocations are reduced (*). The result is a soft, workable state that will stay that way. So, for example if you get 2024-O, it is annealed, and will stay soft and workable. Once it is formed into a part, it can then be solution treated and age-hardened, becoming 2024-T4. (T3 reaches the same state, but is an accelerated aging rather than natural aging)
(*) - I have always found this paradoxical -- that yielding is the result of movement of dislocations, which take far less force than simultaneously shearing a whole crystal along a lattice plane. So more dislocations should allow for more easy yielding. Yet crystals with an absence of dislocations, by being annealed, are softer. I think the yielding process in this case involves the creation of new dislocations.
In principle this process is fully reversible and can be repeated multiple times. However, each time the alloy is heated, or the longer it is heated, the crystals, or grains, can grow larger by absorbing adjoining crystals. Larger grain size can be associated with reduction of some properties, especially fatigue resistance. So it is not a good idea to do this multiple times to something. But doing it once to rivets is fine.