[CliffordK (OP)]

Is there an easy (cheap) way to identify aluminum alloys?  There are a couple dozen common aluminum alloys.  Some weld easy, some don't.  Some can be hardened, some can't.  Some are harder than others.  And, each requires different filler alloys.

Is there an easy way to identify alloys from scrap?

One method is using XRF (X-Ray Fluorescence).  Unfortunately the scanners cost $10,000+, and is somewhat out of the range of a home user.

I suppose I could safely file off a small sample.  IR spectroscopy?  Gas Chromatography? 
Perhaps it would help to just try to determine the melting point and density if I had a good table of physical properties.

[William McCormick]

Is there an easy (cheap) way to identify aluminum alloys?  There are a couple dozen common aluminum alloys.  Some weld easy, some don't.  Some can be hardened, some can't.  Some are harder than others.  And, each requires different filler alloys.

Is there an easy way to identify alloys from scrap?

One method is using XRF (X-Ray Fluorescence).  Unfortunately the scanners cost $10,000+, and is somewhat out of the range of a home user.

I suppose I could safely file off a small sample.  IR spectroscopy?  Gas Chromatography? 
Perhaps it would help to just try to determine the melting point and density if I had a good table of physical properties.

All aluminum mechanically hardens.

If it is common pipe, tubing or extrusion, there are certain probabilities as to what it is.

If you have pipe and you cannot bend it, it is usually 60-61 T6 But could be 6063-T6

If you have pipe and you can bend it is usually 60-63 T-42 or T52
Today common extrusions are 60-63 T-52 60-61 T-6 Both weld well with 40-43 the most common filler wire.

Today 5086 what they call marine grade aluminum is very popular. But marine aluminum comes and goes in and out of favor, as the engineers find out in the real world that aluminum is probably not the material for salt water. Years ago 20-24 was marine aluminum. Now it is 5086. Before that it was 60-61

All aluminum age hardens, so you have to watch out,  a T-3 or T52 hardness today might be something else five years down the road or less. So although you get away with bending it new, it will not bend old.

70-75 tooling plate of old goes as hard as T-8, it will actually stand up to brass. Today though it is more standard to find, T-7 readily available.

The other problem is that 60-61-0 material, can be heat treated, and hardened by local shops. So you never know for sure.

20-24 is a pain in the butt. Because it does not weld right with 40-43 filler, it is like a disease if it is not marked well after you cutup the sheet.

Not really anyway I know of for sure to tell the difference.

For anything important we buy fresh material, and fresh welding rod.

Plus we limit or isolate strange materials in the shop.

Under products you might see what is common right now, and back probably about 20 years. 

http://www.yarde.com/products.html [Links inactive - To make links active and clickable, login or click here to register]

                      Sincerely,

                            William McCormick 

[damocles]

Clifford, I think that your answer might well lie with flame photometry or atomic absorption spectroscopy. Either of these closely related techniques is destructive, but would require only about 1 mg of alloy for sampling. An AA spectrometer is part of the standard equipment of any analytical lab; a cheap but decent commercial instrument costs about a low 5-figure sum. A flame photometer is an easily improvised piece of home hobbyist's equipment.
Either technique works well for aluminium metal.

[CliffordK (OP)]

Thanks, William and damocles.

I had some projects that I wanted to make from used aluminum, but am quickly learning that not all aluminum is the same. 
I'll probably take William's advice that any parts that are extremely critical, and difficult to monitor should be made from new or otherwise "known" aluminum alloys.

Any lab equipment that is 10+ years old will show up on E-Bay and government auctions for pennies on the dollar.  But, may be broken, with limited manufacturer's support, and etc. 

I see there is about a 50 year old flame photometer on E-Bay now.  Perhaps I would have a chance of getting it working and it would be good to practice with.

For a qualitative analysis, perhaps I should just try burning some shavings with my torch and see what happens. 

[RD]

cheap spectroscope thread ... http://www.thenakedscientists.com/forum/index.php?topic=20679

[CliffordK (OP)]

cheap spectroscope thread ... http://www.thenakedscientists.com/forum/index.php?topic=20679

Thanks,
So, all I have to do is figure out how to build the datalogger, and I should be able to come up with a computer interface for a very simple flame photometer/spectrometer.  Time to dig out my "Basic Stamp" interface again.

Then if I had samples of known types of aluminum, then it should be easy enough to compare them with the unknowns. 

Would anodizing be a problem (assuming, of course, that filing it isn't a problem)?

[Bored chemist]

At least some of the elements (Zn, Mg, Si) you would be looking for in aluminium won't work well in a simple flame photometer.
Si is a particular problem since it isn't easy to dissolve in acids unless you use HF which isn't nice stuff.

[CliffordK (OP)]

At least some of the elements (Zn, Mg, Si) you would be looking for in aluminium won't work well in a simple flame photometer.
Si is a particular problem since it isn't easy to dissolve in acids unless you use HF which isn't nice stuff.

Don't they all burn?
I suppose brass alloys can get quite hot without burning.  But, I would think that aluminum alloys would burn better.  Does it have to be dissolved into a liquid solution?

[Bored chemist]

This might help you.
http://www.standardbase.com/tech/FinalHUTechFlame.pdf [Links inactive - To make links active and clickable, login or click here to register]

[CliffordK (OP)]

Thanks,

So the flame photometer is designed for use with a liquid sample, although I'm not sure it would preclude the use of a solid sample.  It would just have to have a completely different feed mechanism and perhaps image capture method.

Looking at the different http://en.wikipedia.org/wiki/Aluminium_alloy#Wrought_alloys [Links inactive - To make links active and clickable, login or click here to register], perhaps a qualitative test would be insufficient, and one would need a quite accurate quantitative test.

What determines the arc color spectrum?  The materials?  or the shielding gas?  Or both?

[CliffordK (OP)]

There are notes on the internet on how to build low-cost X-ray sources.

Perhaps if I could build a spectrometer, then I should look at the X-ray sources as I would imagine the XRF would produce the cleanest, image, and of course, it is non-destructive.

Let's see how much ambition I can muster.

[Bored chemist]

Don't forget that you may need to know the temper as well.

XRF is good, but ambitious.
A spark emission system might be more practical. Charge a fairly big capacitor then discharge it through a tungsten or carbon electrode directly onto the bit of metal you want to analyse. Take a picture of the spark through a direct vision spectroscope with a digital camera.
Compare the picture of the spectrum with a library of spark spectra from known alloys.
I think the problem would be that al lot of the lines you want to look at are in the UV.
Also, you would probably need to produce the spark in an inert gas like Ar to keep the background spectrum simpler and to prevent problems with molecular emissions.
 It's clear why the commercial systems are expensive.

[damocles]

Clifford I have been looking at the wikipedia article on aluminium alloys at
http://en.wikipedia.org/wiki/Aluminium_alloy [Links inactive - To make links active and clickable, login or click here to register]
and it occurred to me that it is going to be very difficult to distinguish the different alloys on the basis of chemical composition with simple equipment. I learnt that aluminium alloys typically contain 95-99% aluminium, for example. Compositions differ by only parts per thousand, and, as BC has pointed out, other aspects of the preparation of an alloy (which affect crystal size, and the concentrations of defects and dislocations) are just as important as composition anyway.

But in trying to look at chemical composition, it might be an interesting idea to digest the aluminium, and filter and wash any residual metal, which could then be assayed/analysed much more accurately, perhaps by wet chemical means.

Aluminium metal is particularly suited to this sort of approach because of the amphoteric nature.

Fraction 1: Digest a fairly large sample of aluminium alloy in hot caustic soda, filter off the remaining metallic solid, and wash with another fraction of hot caustic.

Of the common alloying metals, magnesium, copper, iron, manganese and nickel should not dissolve, and should be available for collection in the residue -- either as metals or oxides.

Fraction 2: Dissolve in warm concentrated hydrochloric acid, followed up with a similar filter and wash exercise.
Silicon/silica will not dissolve, and copper might not totally dissolve. Do not use any acid other than hydrochloric because aluminium metal or alloy will become passive. All of the other commonly used metals will dissolve with the aluminium.

If these procedures are used, then aluminium will be separated from all commonly used alloying elements except titanium, chromium, and zinc, and samples free of aluminium will be available for analysis for the various alloying metals.

The residues of Fractions 1 and 2 should be washed with distilled water and further analysed for the metals they might contain, using either spectroscopic, colorimetric, or wet chemical means.

[Bored chemist]

Damocles' suggestion is a good start, but you need to worry about zinc in some alloys and it's amphoteric too.

[CliffordK (OP)]

One might get a good start on structural aluminum if one could just make the major classifications, in which case one might be able to use a fairly crude qualitative approach.

1000 series (99% aluminum)
2000 series (3 - 7% copper)
3000 series (1% manganese)
4000 series (4.5 to 6% silicon)
5000 series (2 to 6% magnesium)
6000 series   About 1% of silicon, iron, and magnesium.
7000 series (1 to 7% zinc)
8000 series (up to 2.5% lithium).  (I probably won't see much of this, unless somehow I get government surplus rocket components which I probably won't find in the local scrap yard).

Cast alloys might complicate the issue somewhat, but perhaps it could still be lumped with the above.

It also turns out that there are relatively fewer types of welding rods than types of aluminum.  So, perhaps the exact material identification isn't necessary, at least for some things.  I presume part of the issue is that certain more volatile elements cause problems with the welding rods, so they may be in the extruded aluminum, but not in the rods (or final welds).

[CliffordK (OP)]

Don't forget that you may need to know the temper as well.

If welding, I will probably re-temper small parts as the temper could be lost.  I just a good deal on a used kiln.  Some of the controls are bad, but I can get a good digital thermostat on E-Bay for quite cheap, effectively making myself a new digital kiln.

I'll have to decide how I wish to deal with larger parts as the big problem is that the welding process removes the temper around the weld.  Not a huge difference, but it does reduce the strength by a few percent.

A spark emission system might be more practical. Charge a fairly big capacitor then discharge it through a tungsten or carbon electrode directly onto the bit of metal you want to analyse. Take a picture of the spark through a direct vision spectroscope with a digital camera.
Compare the picture of the spectrum with a library of spark spectra from known alloys.
I think the problem would be that al lot of the lines you want to look at are in the UV.
Also, you would probably need to produce the spark in an inert gas like Ar to keep the background spectrum simpler and to prevent problems with molecular emissions.

The capacitor is a good idea, although I can generate spark for a longer period with the welder. 
I think I have argon/CO₂ mixed gas, but may need a different mix for the aluminum welding anyway.
Depending on the resolution of the spectrometer, UV may not be a problem, although X-rays may not behave well with the prisms.

It's clear why the commercial systems are expensive.

Yes & No.
As I mentioned, a lot of older lab equipment gets sold cheaply.  I think the XRF is relatively new, especially with the hand-held devices, so it won't be hitting the used market yet.
I think the actual device, being the size of a barcode scanner must be relatively inexpensive to produce, but it requires some R&D, and as of yet, the market is still limited.
Still, if they made the device for $200, then they might be able to sell it to every machine shop around the world.

Consider how complicated your typical cell phone is.  I suppose one often doesn't see the full price with all the subsidies, but one can find 6 month old phones for dirt cheap.

Anyway, I suppose it is time to start building a photometer.

[Bored chemist]

Have fun, but it seems our ideas of "easy" don't tally.
A spark, rather than an arc is likely to have more of the actual metal in it.