Are you trying to manufacture complex assemblies for use in ultra high vacuum (UHV)? Are you frustrated by the typical prohibition on using anything other than 304 series stainless steel in your chamber? Switch to cheap, light, highly machinable aluminum now, with confidence that it can be properly prepared for vacuum according to the procedure here described. In fact, let's just cut to the chase and present the procedure right now!

Etch for 60 seconds in 5.5% sodium hydroxide (NaOH) at 70 C.
Thoroughly rinse in tap water.
De-smut for five minutes in 25% nitric acid (HNO3) and 0.1% ammonium bifluoride (NH4HF2) at 20 C.
Transfer to DI water for no more than ten or fifteen minutes.
Rinse in DI or hot DI water if available.
Completely dry with filtered compressed air or even better, with nitrogen or hot nitrogen.

In following this procedure, you will be following similar practices to those used at CERN, the European Organization for Nuclear Research which operates the Large Hadron Collider, and SLAC, the Stanford Linear Accelerator, both of whose aluminum cleaning procedures were reported in this survey commissioned by the American Vacuum Society.

Although the procedure here is highly similar in underlying principles to those used at CERN and SLAC, the exact concentrations I chose were set by the practices that were being used in the anodization process here at JILA, the Joint Institute for Laboratory Astrophysics, a joint venture of Colorado University and NIST, the National Institute of Standards and Technology. Anodization begins with etching and cleaning, just like vacuum preparation.

By sticking with the same concentrations used in the anodization line here, I ensure the later usefulness of the chemicals. Besides, some of the concentrations used at large labs are rather extreme, due to their need to perform the cleaning by portable sprayers because the components don't fit in baths.

Step 1: So What's Really Going On?

Usually, we are taught to fear aluminum due to its oxide layer. Aluminum is highly reactive with oxygen, so much so that once extruded in a forge, the aluminum rapidly forms an oxide layer as it cools. Unlike iron, which slowly forms an oxide that then crumbles away and exposes more iron to oxidation, aluminum's oxide can be dense, strong, and protective of the metal within from further oxidation.

Unfortunately, by UHV standards, the aluminum oxide layer is usually thick (tens of nanometers), dirty, porous, and difficult to clean, especially when the oxide is formed during extrusion of molten aluminum in a mill. Even when formed during cold machining in the shop, oils and other contaminants can be trapped in the oxide layer. Worse still, ultrasonic cleaning procedures, the workhouse of ultra high vacuum material preparation, are too aggressive for aluminum and can cause pitting and increase porosity of the oxide layer.

The solution? Chemically etch a few microns away from the Aluminum surface, completely stripping the existing oxide layer formed during machining or extruding. The new oxide layer that is instantly formed after the etch will be incredibly dense, thin (angstroms), and nonporous, since it is formed at cold uniform temperature and low pressure directly in the atmosphere's oxygen. After a few cleaning steps, the aluminum is ready for UHV, and since the oxide layer is permanent there is no rush to get the parts into vacuum.

In fact, these days vacuum chambers are actually being made from aluminum, and lower pressures are being obtained compared with steel. This company is selling aluminum chambers, and here are their arguments for why it is a much better choice than steel.

All of the information I just provided was gleaned from discussions with Chemists and Machinists, from the very useful commentary beginning on the bottom of page 2031 of the same survey article I linked in the Intro, and from various informational blurbs on aluminum processing company's websites.

Step 2: Materials

40 Liter flat-bottomed Steel Tank.
Large hotplate, Corning PC620D (1.1kW) with teflon probe. Manual.
Extra thermocouple probe and reader.
20 Liter "carboys" for cleanup, x3.
40 Liter secondary containment bin.
Large blocks to take the weight from the hotplate.
1kg Sodium Hydroxide (Lye) pellets.
5 Liters of concentrated 70% Nitric acid.
25g of ammonium bifluoride etching powder. Order from Sigma Aldrich.
Fume Hood.
DI Water Tap.
Nitrogen or Compressed Air blow nozzle.
Additional large bin or container for DI water dip.
Nitrile gloves and shoulder length rubber gloves for cleanup.
Measuring beaker, at least 1 Liter.
Large Funnel.

The chemicals used here are standard inventory for any chemical supplier, with the possible exception of the etching powder, although that is commonly used for etching marks on glass, which might make it more readily available than you'd otherwise think. The large stainless steel tank might be hard to find unless you have a shop nearby that does similar things already, as in my case. There's probably a safer choice of glove than nitrile, but less intense than the shoulder length rubber. I just used nitrile most of the time and should length rubber when I thought there was a chance I could get splashed during cleanup.

Step 3: Prepare the Etchant Bath

Start by preparing the etchant bath, since it will take a long time to heat up. Mount the 40 Liter stainless steel bin above the hot plate in such a way that the hot plate is firmly touching the bin but not bearing more than 25 pounds or so. I achieved this using some large aluminum blocks I found in the shop-room stock area. The blocks were cut from the same extrusion so they worked perfectly for providing an elevated but flat mounting structure. Make sure you place the bin on the floor close to the front of a fume hood. It won't fit inside, but this way the fumes can be exhausted through the hood simply by opening the front.

Carefully measure in 35 Liters of tap water. There's no need for cleaner water since any impurities will be completely dwarfed by the compounds we add. I only had a 1 Liter beaker on hand, so this took fifteen minutes. If you can find a larger measured vessel, this would save time. For that matter, just measure five liters into a larger vessel and label it with a sharpie, now you have a larger measured vessel. Wish I thought of that before!

Before rigging up the heating plate's temperature probe, just fire the thing up to full blast while you go on to the next steps. At 4kJ/kgC, it would take 140s/C*45C = 1 3/4 hours to increase the bath to the target of 70C without any losses to convection with a 1kW heater.

Now mix in 0.973kg of sodium hydroxide pellets, so as to obtain a 1.39 mol/L or 5.5% solution. Sodium hydroxide (40g/mol) is a powerful, caustic base so wear gloves and don't splash. Be sure to weigh out the pellets under a fume hood, because the dust that comes off while you are pouring the pellets will aggravate your lungs. (trust me, I learned this the hard way. It didn't hurt really but just made me catch my breath in my throat which was maybe slightly sore that night). The sodium hydroxide won't fully dissolve until you raise the temperature of the bath, but it's not a very exothermic reaction, so you don't have to worry about splashing or bubbling when it heats up and dissolves later.

Now go back and rig up the teflon probe. I did it with some chemistry mounting frames I found lying around. Turn the hot plate off and back on after plugging in the probe, and set it for 70C. You might want to track down a thermocouple, since the Corning Style hot plates won't tell you what the temperature currently is, only whether it is at target or not. You'll also need to stir occasionally, or a large temperature gradient will develop. Make sure the bath is mostly covered, and move on to the next step.

Step 4: Prepare the De-Smut Bath

We will use the secondary containment bin as a primary container for our acid de-smutting solution. This is acceptable in the short term- secondary containment is only required for longer term storage.

Begin by adding 12.5 Liters of water to the bin, and then slowly add 5 Liters of 70% Nitric acid, so as to obtain a 20% concentrated bath. (actually, I didn't account for the increased density of the acid, so this actually gives a 25.3% solution) This is a very exothermic dilution, so it is very important to pour slowly to avoid locally boiling the water and splashing the acid. It is also important to avoid heating the bath beyond 80 degrees C, the softening point for the high density polyethylene of the secondary containment bin. I poured slowly enough so as to have added a 2.5 Liter jug after five full minutes. At the beginning when the dilution is most intense, I could see tendrils of steam coming off the water as I poured.

It is a good idea to monitor the temperature as you pour. After adding half the acid, my bath had risen above 30C, and at the end it had reached 40C. Actually, lets check this:

70% HNO3, is 70% by weight, so there are 70g HNO3 in 100g solution. The density is 1.413, so there are .7*1.413 = 989 gm/L, or 15.7mol/L after dividing by the 63g/mol mass of HNO3. Also, 70g in 30g water means the molality is 37mol/kg. Now, after adding 12.5 Liters of water, we have 989*5 = 4.945kg HNO3 in 2.12+12.5 = 14.62 kg of water. This is a 5.37 m (mol/kg) solution, or 25.3%. Looking in this table, we find that the new density is 1.15 gm/L.

Now, looking up the heat of dilution, I find it should be something like 15kJ/mol, in between the top two columns of the table. We're not diluting to infinity, just to 25.3% (5.37 m), where the heat of dilution is 1.4kJ/mol. That means we should release 13.6kJ/mol of acid, and we have 78.5 mol, so 1068kJ. Now the water should go up a degree C for every 4kJ/kg*17.5kg = 70kJ, or 15C. Of course, the specific heat is different for the acid, 1.7kJ/kgC, but this is a small correction: 14.6*4 + 4.95*1.7 = 66.8kJ per C, so 16C from 1068kJ instead of 15C.

Hooray, my final temperature measurement of 40C is consistent with an initial temperature of 24C, a bit warm for tap water but within the realm of possibility. Chemistry works!

Step 5: Spike With Fluorides!

Now we add the ammonium bifluoride. It was exciting to unwrap the multiply packaged chemical from sigma aldrich. I could have used the less pure grade, but I wanted to see what the chemical looked like in pure form, and its cost was chump change relative to my experiment budget ($70). The chemical was a sticky, flaky mass almost like the really big snowflakes you can find during a flurry at altitude (no joke, I've seen flurries entirely comprised of 3-5mm complete flakes here in Boulder, CO. Before moving here from the east coast, I thought the whole paper snowflake was just an artsy thing. It's not, they really look like that).

The dilution is also very exothermic, about 20kJ/mol if you scroll way down that link, but with only 25g we're talking 7.3kJ. This means it is safe to dump it right into the giant bath without any heating trouble, but the chemical is extremely dangerous otherwise. If you follow the link it talks about fluoride ions causing anything from burning to cardiac failure to death. We have to have fluorides in the de-smut bath so that it can remove silicon from the smut that forms after etching. Silicon will be present because it is almost always a minority component of aluminum alloys. I used a much smaller quantity of fluorides than the national labs, because I knew my aluminum grade (6061) was much lower in silicon than some of the others. At CERN they used hydrofluoric acid for their fluorides, but ammonium bifluoride is slightly safer to work with- (maybe because it is solid at room temp, I'm not sure).

Step 6: Final Preparations

Find some steel welding wire and rig up a way to dip your parts in the etchant and the de-smut. If your parts are going to be useful for anything, they must have some kind of screw hole or other feature you can use for this. If you have to etch perfect spheres or something, good luck. Maybe just rig up a holder like they have for dying easter eggs. But if your part gets stuck in the etchant, you're in trouble. The etchant will completely eat up the aluminum until the solution is no longer basic, at which point you'll have to add a lot more sodium hydroxide pellets or safely dispose of a lot of waste and start over.

Position a DI water bath very close to the de-smutting acid, so that you can put parts directly into DI after their acid bath, and not carry them over to the sink, dripping acid the whole way. You might want to do the same for your etchant bath, since you'll be rinsing in between etching and de-smutting. I didn't do this because my etch bath was right next to the sink in the fume hood.

Do a dry run to make sure you have a good way to time everything properly. I used a lab timer like the one shown, which is awesome because it begins counting up once it finishes counting down so that you know how long it has been since the timer ran out.

Etch for exactly one minute
Thoroughly rinse
De-smut for about five minutes

Step 7: Perform the Etch!

Now it's time to actually perform the procedure. When you dip in the etch, there will be a few brief moments where nothing happens, maybe because the solution hasn't eaten through the oxide layer yet, or because the aluminum has to heat up a bit. Soon later however, an intense bubbling will begin. Make sure you move the part around to help the etch perform evenly. My first part came out with some marks, and I think it may be because I didn't do this.

When one minute is up, lift out the part. It will be ugly looking and will steam threateningly! Rinse it thoroughly in the tap before moving on to the acid dip.

Step 8: De-Smutting and Cleaning

Start your timer for the five minute de-smutting acid dip. Within ten seconds, you'll see the part mostly become clean in the acid, but be patient in case any minority components are slower to come off, and move the part around occasionally. Transfer to your DI bucket after the time is up.

Be careful to make sure that the attachment you use for dipping is dipped further into the rinsing DI water than into the acid. Otherwise there will be acid remaining on the attachment to burn you or mess up your parts.

After the dip, rinse very thoroughly in DI water, just make yourself keep doing it for a long time. Most DI taps are somewhat slow, so be patient.

NOW MAKE SURE TO COMPLETELY DRY YOUR PARTS! For some of my parts, I didn't do this, because the Nitrogen tank ran out, the filtered compressed air line was temporarily down, and I didn't feel like fixing either. I sorely regret this, because the few drips of DI sitting on the parts severely discolored them, presumably damaging the thin dense oxide layer I was seeking to create.

Step 9: Cleanup and Waste Management

Transfer the 35 Liter etch bath to two 20 Liter waste carboys. Make sure they are HDPE, so the 70C liquid is safe since HDPE only just begins to soften above 80C. Don't try and let the bath cool force or the sodium hydroxide will precipitate out as salt. You don't want to touch that salt or have to scrape or clean it out. Make sure you use a large funnel and pour slowly. I wore shoulder length gloves for this part, as it has the highest propensity for spilling or splashing.

Transfer the 17.5 Liter acid bath to the remaining 20 Liter waste carboy in the same way and also wearing shoulder length gloves. Rinse out the baths and then use the acid bath for secondary containment. In my case, these chemicals will soon be permanently installed as part of an aluminum anodization line, since anodization is often preceded by etching and de-smutting.