Despite the fact that very few people actually have access to equipment to do this, I thought I might share some of my experiences on the aluminizing process. At any rate, most ATM's are of the tinkering sort and should at least be interested in knowing how this is done. I might point out that this process is described in "Amateur Telescope Making", Volume 3, Albert G. Ingalls ed., Willmann-Bell Inc. Another good (and almost as old) source for information on aluminizing as well as many other interesting technical curiosities, is John Strong (et al.)'s, "Procedures in Experimental Physics", Prentice-Hall, 1938. While we're on this topic, John Strong's, "Procedures in Applied Optics", (M. Dekker Inc., New York, 1989) is also an interesting read for ATM's... if you can find it. You will probably have to dig into a university library to find these books. The information presented here assumes some familiarity with the process as described in these or other books.
The equipment required to evaporate thin films of metal on glass or other suitable substrates is standard to most university physics research labs. Although the equipment is only moderately expensive (as far as physics equipment is concerned) it's well out of reach of most individuals, especially in view of the low cost and high quality of commercially produced coatings. Nevertheless, let's assume for now that for some reason or another you have access to a high vacuum evaporator; here's a few things I noted.
I never tried evaporating aluminum above a pressure of 5x10-6 Torr so I can't be of help on this. 10-4 Torr is considered to be an upper minimum for the required pressure. You'll need a diffusion pump to get a pressure this low; a diffusion pump can be improvised from stainless steel if you're a very good welder. The pump oil though is very hard to improvise; this must be bought (at substantial cost) from a high vacuum supplier. The glass mercury pumps you may have seen discussed in older articles are an absolute no-no, even if you manage to get your hands on enough mercury; mercury pumps are gone and best forgotten. Glass diffusion pumps won't be of much use anyway because they are too difficult to construct in the large size required. Good pumping speed is essential or stray diffusion pump oil will eventually find its way onto the glass being coated. Essentially, the glass should be cleaned and coated in as little time as the system takes to reach the required deposition pressure. Liquid nitrogen cooling traps largely solve this problem, but are not strictly necessary if the pumping speed is fast enough (i.e. less than an hour to reach 10-5 Torr after the diffusion pump is turned on).
O.K., so now you need something to evaporate - this is the easy part. It turns out that aluminum foil (i.e. the baked potato kind) works just fine thank you. I received quite a few sceptical looks from colleagues who are used to using reagent grade aluminum in their evaporations when I told them about this. I suspect that unless you get the glass REALLY clean, the impurities in the commercial grade aluminum foil won't make much difference to the adhering of the aluminum to the glass. I found that quite a bit of aluminum foil is required to get an even coat over an 8" mirror. The foil was folded over several times before it was wrapped around the tungsten heater coil.
The crux of the operation (assuming that you can get the vacuum system to a low enough pressure) is in cleaning the glass. The glass must be absolutely chemically clean; this means soaking the glass in ~10% nitric acid for a half hour or so followed by soaking in chromic acid. I don't have access to chromic acid for my evaporations so I just do without.
The water must then somehow be chased from the glass before it has a chance to evaporate and leave water marks on the surface; a very frustrating endeavour. I have often read about the use of a very clean "over laundered" towel to do this, but I have never had success in applying it. This always leaves conspicuous smudges on the surface (KimwipesTM aren't much use either). Instead, I use high purity acetone to chase the water from the mirror surface with the mirror tilted at a very high angle. This also leaves water marks to some extent, but if done carefully I find that these are confined to only the very bottom edge of the tilted mirror. After this, nothing is allowed to touch the mirror and it is immediately inserted into the high vacuum chamber.
The easiest way to effect the final cleaning of the mirror is to apply a very high voltage (see disclaimer below) between the bottom of the evaporation chamber and an electrode located behind the mirror. At relatively high pressures this will produce an electron discharge within the chamber that excites the rarefied molecules of air in the chamber which in turn emit light (the process is very similar to that responsible for the Aurora Borealis/Australis). The ions produced in this process impinge on the glass at great velocity, dislodging residual oil and other molecules from the surface (or so the story goes). For my first coatings I didn't bother with this cleaning step: BIG MISTAKE - the films produced degraded rapidly in the space of a couple of months. For my last set of coatings I used a Tesla coil to produce a glow discharge at a pressure of 10-1 to 10-2 Torr within the chamber. I left the chamber for about 30-40 minutes in this condition before opening the diffusion pump to lower the pressure to the evaporation pressure. The films thus produced were very bright with very few blemishes and have not degraded noticeably in the last five months of use. I suspect however that these films will not stand up to much washing.
It was a pleasant surprise to find that a LOW POWER Tesla coil would produce sufficient discharge to effect the cleaning. This is safer (and cheaper) than using a high power transformer such as a neon sign transformer for the cleaning. The discharge current need not be DC; in fact the output from a Tesla coil is not at all smooth but very "spiky". I emphasize the use of a low current Tesla coil.
Now for the DISCLAIMER: If you happen to build a high power Tesla coil (such as the ones described on various web sites) or set up a high voltage supply using neon sign transformers or any other high voltage device and zap yourself in the process, I will bear no responsibility for your actions.