Sciencemadness Discussion Board

Galinstan-activated aluminum: an inexpensive, easy-to-use, easy-to-store, nontoxic alternative to Al/Hg reduction

Melgar - 18-10-2016 at 16:26

A lot has been written regarding using aluminum/gallium alloys as reducing agents in organic chemistry, or as a source of hydrogen gas when mixed with water. However, the gallium content of these alloys needs to be quite high in order to produce a reaction, which is less than ideal, considering gallium doesn't participate directly in the reaction, and is several orders of magnitude more expensive than commercially-pure (99%+) aluminum.

Galinstan (Gallium + Indium + Tin) is a low-melting-point alloy that melts at a lower temperature than ice, containing gallium as the majority of its makeup. This alloy can be prepared simply by heating 68% (by mass) Ga, 22% In, and 10% Sn. The tin may need elevated temperatures to dissolve, although a Ga/In alloy will still be liquid at room temperature. Commercial galinstan also has small amounts of antimony and bismuth in it (<1.5% wt.) to lower the melting point and improve other physical properties.

Preparation of the galinstan alloy

To 34g of molten gallium (35˚C) was added 11g of indium scrap. The indium readily dissolved, but left behind a significant amount of papery indium oxide that floated on the surface of the alloy. 0.5g of additional indium was added to compensate for the indium oxide that didn't dissolve. The indium oxide was removed and saved for possible use later.

For the tin source, rosin-core lead-free electrical solder was selected, mainly due to it being readily available, but also due to the fact that it contained 5% antimony (the other 95% being tin), which was noted in a German patent to improve the alloy's physical properties. 8g of solder was removed from the spool, and heated with a propane torch in a stainless steel crucible. The rosin was burnt off, until only a thin lump of molten solder was left in the crucible. The lump was weighed, and found to have only lost about 0.5g as a result of burning the solder. 2.5g was cut off the solder lump with wire cutters, leaving 5g that was added to the alloy. Unlike the indium, the tin did not readily dissolve, and so the alloy was heated with a propane torch in a pyrex test tube to about 200˚C, at which point the solder mixed with the rest of the alloy. A solidified droplet of bismuth was added at this point, that had formed earlier when doing an experiment with molten bismuth. The droplet was not weighed. It dissolved readily at the higher temperature. The alloy was allowed to cool, and moved to a separate holding vial. The alloy had wetted the inside of the test tube, and this was removed by adding a very dilute aqueous HCl solution and shaking. The alloy collected in a large droplet at the bottom of the test tube, and was removed with a pipette and added to the rest of the metal. (Aside: I think the reason this works is that the HCl creates like charges on the surfaces of both the glass and the metal, which subsequently repel each other. This also works with dilute NaOH, though not quite as well.)

Preparation of the aluminum/galinstan alloy

For the aluminum source, I used dollar-store aluminum cooking pots that had clearly been stamped out of sheets. 1100 aluminum alloy works particularly well for stamp-pressing, and is 99% aluminum, which accounts for its softness. Aluminum electrical wire also works, although the preferred alloy these days is usually only 98.5% pure. Casting aluminum and structural aluminum are the least pure aluminum alloys, and should be avoided. The two main impurities in aluminum are iron and silicon, which can be tested for by dissolving a small sample thoroughly in clear half-strength hydrochloric acid. Yellow color indicates iron, while an insoluble light brown precipitate indicates the presence of silicon.

For melting the aluminum, you will need some sort of crucible. I bought mine from a jewelers' supply shop for $5, but it only needs to be able to withstand 600-800˚C, so material is not that critical. You can use an indentation in concrete or a large enough rock as a crucible if you have to, since you won't actually be pouring any molten metal. These would be better than using a steel crucible, which would contaminate the melt to an unacceptable level. You'll need something to stir with; graphite rods are ideal, but if you're doing this outside and smoke isn't a problem, you can use a stick. It will gradually burn away, but will only form charcoal and ash, neither of which will contaminate the aluminum. Really, just don't use anything metal for stirring unless you know what you're doing. You'll also need a heat source that can apply enough heat to keep all your aluminum molten at once. A propane hand torch with a one-pound tank like the one pictured can keep about 50g of aluminum molten at once. For reference, this is 50g of melted-down aluminum:

files.php.jpg - 79kB

If you don't know how much aluminum you'll be able to melt at the start, you can cut up your source into pieces and do something like this to keep track of how much metal you're melting:

files.php.jpg - 160kB

Start melting down aluminum, until you have the desired amount of molten metal, or until you reach the limit of how much metal your heat source can keep molten at once. If necessary, you can cool it down, weigh it, then melt it again. Now weigh out an amount of galinstan equal to about 5% the mass of your aluminum. Smaller amounts can work, just not as consistently. Larger percentages of galinstan will cause the aluminum to react more quickly. After much trial and error though, I've settled on 5% as the best balance between allowing a smooth, predictable reaction and keeping galinstan use low due to its relative expense compared to aluminum. You can weigh the galinstan in paper (preferably low-ash filter paper) then set it all into the molten aluminum to minimize splatter. As soon as the two mix, you'll notice the consistency change. The melting point drops, and it's easier to keep molten. You'll also notice that it forms a cement-like phase before it solidifies, unlike pure aluminum. It is critical at this stage to throughly mix the two metals. Once the metals are thoroughly mixed, the molten metal can be covered and allowed to cool to below 100˚C. At this point, the metal should have the approximate strength and physical properties of unfired clay; that is, not much strength at all, and the metal should be easy to break using two pairs of pliers to grip it:

files.php.jpg - 111kB

If the metal is too hard to break in spots, that's usually a sign that those spots didn't get enough galinstan mixed into them and will probably need to be remelted with more galinstan in order to make them useful as a reductive agent. This is a good time to break up the metal into small pieces and store it, preferably in an airtight glass jar. When added to water or alcohol, the metal will gradually break up into grains about the size of sand, so you don't need to manually break it up into very small pieces. This alloy reacts slowly with air, and can be handled with dry hands, even without gloves. The only danger in handling it, is if your hands are wet, and either you touch it or a small piece of it sticks to your skin, it can generate enough heat from the reaction with water to burn you, though not very badly.

Typical reaction conditions

Substrates are initially dissolved in either water, a lower primary alcohol, or a mixture thereof, which act as hydrogen donors for the reduction. Isopropanol can work, however under anhydrous conditions, it will react with aluminum to form aluminum isopropoxide, which can reduce ketones into alcohols, and transform aldehydes into esters.

For reductions in methanol, ethanol, and water, nitroalkanes are reduced to amines in excellent yields. Nitroolefins are reduced into their corresponding alkyl amines due to a rearrangement that produces an imine. Reductive amination also proceeds smoothly. Stirring is highly recommended, since this is a heterogeneous reagent, although in my experience, the alloy usually generates enough hydrogen bubbles to adequately circulate the solution. Reaction rate can be controlled by the application of heating or cooling.

This reagent is also a convenient hydrogen source when added to water, and can be used as a more reactive form of aluminum in a number of reactions. For example, it can be added to a solution of iodine in diethyl ether to give an ether solution of aluminum triiodide. It can also be reacted with isopropyl alcohol to give aluminum isopropoxide, which can catalyze the conversion of ketones to alcohols in the presence of isopropanol, and secondary alcohols to ketones in the presence of acetone.

galinstan_scale.jpg - 125kB aluminum_melt.jpg - 175kB breaking_up_blob.jpg - 111kB reaction_with_water.jpg - 18kB

[Edited on 10/19/16 by Melgar]

aluminum_blob.jpg - 79kBaluminum_pieces.jpg - 160kB

careysub - 19-10-2016 at 11:42

No one has commented publicly yet, so I just want to say that I am glad to see this write up (indeed, I requested that one be prepared on another thread).

Soluble mercury compounds are very hazardous materials, so a procedure that dispenses with them is a major step forward in green chemistry!

brubei - 19-10-2016 at 12:18

Nice but it needs some sources.

Maroboduus - 19-10-2016 at 15:18

So, is the 'cement like phase' the aluminum coming out of solution?
Is the final result something like aluminum in a Galinstan matrix?

careysub - 19-10-2016 at 15:56

Quote: Originally posted by Maroboduus  
So, is the 'cement like phase' the aluminum coming out of solution?
Is the final result something like aluminum in a Galinstan matrix?

I believe what is being described is the formation of a non-ductile mass of aluminum, due to gallium segregating at grain boundaries and causing embrittlement.

So the aluminum is simply solidifying and forming crystals, and the gallium is coming out of solution along grain boundaries.

This is a well known phenomenon with aluminum and gallium, see for example this paper, "Brittle-to-ductile transition in polycrystalline aluminum containing gallium in the grain boundaries".

This work is producing an alloy essentially identical with the 95/5 alloy mentioned here (and their conference paper is also attached).

Attachment: woodall2008.pdf (519kB)
This file has been downloaded 159 times

[Edited on 19-10-2016 by careysub]

Attachment: itoh2010.pdf (3.2MB)
This file has been downloaded 180 times

careysub - 19-10-2016 at 15:59

See also this presentation:

Attachment: production_woodallziebarth.pdf (643kB)
This file has been downloaded 157 times

Melgar - 20-10-2016 at 00:54

Quote: Originally posted by careysub  
See also this presentation:

Wow, they figured out a lot of the same stuff I did, but with better equipment and a better understanding of chemistry.

They also refer to the "Al-Ga-In-Sn" system. I was going to initially call this alloy "Algalinstan", but then decided that "Galinstan-activated aluminum" sounded more professional.

As far as references, believe it or not, I never used any. I think back in 2009 I heard about aluminum and gallium forming a hydrogen-generating alloy, but with the downside of requiring a lot of gallium. You couldn't get gallium quite as easily then, so I used galinstan from a mercury-free thermometer. I also had a bunch of aluminum Chilean one-peso coins, each of which was worth 1/5 of a cent, and I found that if I put a drop of galinstan on one, and scratched at the coin through the drop with a knife, I could get them to mix, which would make the metal extremely reactive with water. I found a reference on the Hive or Rhodium or one of those sites about reductive amination with aluminum dissolved in gallium, and found that reductive amination also worked with what I had. It would also reduce nitromethane to methylamine, which could be detected very easily by smell. When I finally was able to get pure gallium, I found that alloys of aluminum and gallium were nowhere near as reactive as aluminum and galinstan, so I just played around with ratios until I found one that worked consistently, and was slow enough that it approximated the reaction speed of an Al/Hg reduction. This was back in 2009 or 2010. I posted about it, but never did a detailed writeup, just experimented a lot with it.

Melgar - 28-11-2016 at 08:36

Just for the record, I found this page, which indicates the specific reactions that Al/Hg amalgam undergoes:

I imagine this alloy behaves similarly, unless any of the metals catalyze additional reduction reactions. One thing that that page doesn't mention is that after reducing a nitroolefin to an amino-olefin (enamine), the enamine rearranges into an imine, which is easily reduced under these conditions. (Vogel, 5th ed. 769) However, there can be side reactions that produce dimerization before it gets to that point.

Melgar - 10-1-2017 at 15:10

Another thing I'm discovering as I do more experiments with this reaction, is that aluminum purity seems to be very important for determining reactivity. Aluminum alloys typically only vary by 1-2% for metal content, but the reactivities of those alloys vary wildly. Also, I've found that these pans are MUCH more reactive, after being melted down, than any other aluminum source I've used:

pl3996927-takeaway_disposable_aluminum_foil_serving_trays_for_food_323mm_266mm_64mm.jpg - 63kB

Preliminary research points towards 3004 aluminum as the likeliest contender, and the added magnesium could be what's increasing the reactivity. I have yet to determine if or how magnesium changes the reducing activity of the alloy though. I'm having a hard time determining how aluminum reacts with organochlorides and organobromides, though magnesium's reactions with organohalides are extremely well-documented. The only thing that I seem to be able to find are warnings that methylene chloride and chloroform can undergo very energetic reactions in the presence of aluminum, but there's no indication of what the products of that reaction would be. Presumably it'd result in various chlorinated alkanes and alkenes? There may just be no information about the reaction products because the reaction is useless for synthetic purposes.

As far as reducing galinstan content in order to reduce reactivity of the resulting alloy, it doesn't work very well below 5% or so. I did 97% aluminum, but then the galinstan doesn't diffuse throughout the metal uniformly, and different pieces of the resulting alloy vary a lot in their reactivities. Another thing that I intend to try is an alloy composed of a eutectic mixture of just indium and gallium instead of galinstan. That ends up being 3:1 based on mass of gallium to indium. It melts at 15˚C, which is 59˚F. Stannous chloride could then be added to a reaction mixture as needed in order to fine-tune reaction rate. However, it seems to make the most sense to try running the reaction with only group 13 elements (aluminum, gallium, and indium) at some point, to characterize it in its simplest form. Since tin is in group 14, it would seem to be most likely to contribute to variation from the Al/Hg reaction.

One last observation: getting zinc or zinc salts into your reaction mixture is a mess, and I would recommend avoiding it as much as possible. Zinc barely dissolves in galinstan, but is plated onto it if aluminum metal is present. This results in thick, dark gray zinc mud everywhere, which can reduce gallium metal out of solution. Normally, gallium is reduced back onto the dissolving aluminum, and so the amount of gallium in the solution is always quite low. Zinc can interfere with this equilibrium though, pulling more and more gallium out of the reacting alloy until it's no longer able to form a liquid phase and the reaction stops. So to anyone thinking about trying to use zinc dust to reclaim galinstan: it wouldnt' work, and would probably make your job a lot harder. Instead, try this:

To the collected and washed sludge from a dissolving aluminum reaction, slowly add 20% HCl solution until it becomes clear or stops dissolving. Filter, then slowly add small pieces of aluminum (NOT dust!) to the liquid, waiting a few minutes after each piece to see how vigorously it reacts. Hydrogen bubbles should evolve, and eventually, parts of the surface should form a liquid phase. The fraction of the metal in the liquid phase should gradually increase, eventually getting you back your galinstan. This galinstan will be less pure though, due to impurities from the aluminum that dissolved during the initial reaction. If you're tempted to try reducing the metals with zinc, just don't. In practice, it dissolves at a similar rate as gallium, and the component metals of galinstan do not combine in a liquid phase like they do when using aluminum. Gallium seems to alloy with the zinc, forming a higher-melting-point alloy than pure gallium, and then the other metals can't combine with gallium to form a liquid-phase alloy. It just results in a mess, without much to show for your efforts.

Googolplex - 22-6-2017 at 06:37

Quote: Originally posted by Melgar  

Larger percentages of galinstan will cause the aluminum to react more quickly. After much trial and error though, I've settled on 5% as the best balance between allowing a smooth, predictable reaction and keeping galinstan use low due to its relative expense compared to aluminum.
[Edited on 10/19/16 by Melgar]

If you say "a smooth, predictable reaction". Can you give me a reference, like the a amount of aluminium and galinstan used and their reaction time?

Would it be possible to use the Aluminium and Galinstan as it is, for reduction reactions (in-situ)? What I mean is without melting Al or fusing Al before you gone do the actual reduction reactions. Its should still be liquid beneed room temperature. And therefore I suppose it still will attack the aluminium in the MeOH, I dont know if i'm right

Thought experiment that I thinking about :
In a three-neck round bottom flask fixed with a condenser I put 27.5g aluminium (99.995%) granules (2 - 3mm), 750 mL of lab grade MeOH. In a a dropping funnel I mix 25g of ketone, 20g of nitromethane with 25 mL of MeOH. Now I drop the Galinstan in the round bottom flask. And wait until the reaction starts. Then open the valve of the dropping funnel as it would be empty in about 40 - 45 min.

Will this work, do nothing or take longer than 45 min for all the aluminium to be comsunt by the Galinstan? I'm thinking this way because I think the Al/Galinstan reduction reaction is to vigorous and would result in a short reaction time and temperatures above 50°C. It's very important that it goes for 40 - 45 min, without exceeding 50°C.