04 June 2007

The Aluminium-Hydrogen Concept

Awhile ago there was a concept floating around the blogosphere and some alt media such as PhysOrg and Gristmill about burning a Aluminium-Gallium alloy to produce hydrogen gas for a fuel cell vehicle. Apparently incorporating a significant amount of Gallium inhibits the formation of the protective oxide layer that normally prevents aluminium from spontaneously combusting upon contact with the atmosphere. This is apparently the brainchild of Prof. Woodall at Purdue. He has a Powerpoint presentation that details the idea. It probably would have been more worthwhile for me to write about this when Robert Schwartz first prodded me a month ago, but hey better late than never.

Gallium is commonly alloyed with Aluminium, usually in conjunction with Arsenic, because they have nearly identical crystal structure but different electronic and optical properties. This allows one to grow thin layers of material with little built-in strain but rapidly changing electro-optical behaviour. It is this area of physics that is Woodall's background as an electronic engineer.
Figure 1: Gallium-Aluminium phase diagram showing proposed alloy composed of 72 % Gallium by weight [taken from Woodall's presentation].

The first issue I noted is that it takes a lot of Gallium to inhibit protective oxide formation in Al. As shown in Figure 1, more than 2/3rd of the material is Ga and hence does nothing but act as dead weight. Gallium isn't cheap, at something like US$400/kg. Woodall was quoted by Physorg as stating,
"A midsize car with a full tank of aluminum-gallium pellets, which amounts to about 350 pounds of aluminum, could take a 350-mile trip and it would cost $60, assuming the alumina is converted back to aluminum on-site at a nuclear power plant."
350 lbs. of aluminium corresponds to 900 lbs. of Gallium with an approximate value of US$160,000... and people think Li-ion batteries are expensive. Keep in mind that while this Gallium isn't consumed, it also doesn't stay with the car! It has to be traded to some vendor, the AlGa equivalent of a gas station. The potential for fraud and theft is beyond pale. How do you weigh the Ga independently of the alumina? Doesn't trading a $160,000 block of Gallium and Aluminium Oxide to get a $160,060 block of Gallium and Aluminium seem a little kooky?

A problem of similar magnitude with the concept is its tremendously poor cycle efficiency. Consider the enthalpy of the two associated reactions:
Al + O2 → Al2O3 + 1675 kJ/mol

H2 + O2 → H2O + 285.8 kJ/mol
The above two equations aren't balanced, but the enthalpies are correct for the right-hand side. Now the sum reaction would be,
2 Al (bulk) + 3 H2O → 2 Al2O3 + 3 H2 + 2492 kJ/mol
This should immediately raise alarm bells. Of the energy used to produce the aluminium in the first place, only 33 % is going into the hydrogen. The rest is released as heat. All that Gallium, and all that alumina reaction product, are going to act as a thermal buffer, thereby preventing you from using the heat in any sort of useful way. I haven't seen anything from Woodall to suggest one could recover this energy. When you consider that a fuel cell is perhaps 50 % efficient, and electrolysis of aluminium might be 50 % efficient on a good day (compared to 75 % for hydrogen from water), you have a system with a round trip efficiency of less than 8 %. It makes the hydrogen economy look like a paradigm of efficiency in comparison.

There are other major issues, most of which are enough to shoot down this idea on their own. Consider, for example, how do you control the rate of the reaction? If you want to deliver hydrogen to a fuel cell on demand, dripping water into a 150 kg block of AlGa alloy isn't going to get it done. The water will only be able to diffuse slowly through all the previously reacted products. In other words, not only do you need a hefty mass of AlGa pellets to make this concept work, you also need to store those pellets inside a pressure vessel in order to collect and buffer the hydrogen that is reacted. Furthermore, there is no way to refill a tank until it is empty! Hydrogen as a gas is quite difficult to pump to vacuum, which would be required to vent the tank for safety reasons.

It also has major distribution issues. Hydrogen can be pipelined; hydrogen can be reformed from natural gas or electrolysed from water locally. Hydrogen cannot be distributed with existing infrastructure, but it's possible. I can't imagine an aluminium smelter on every corner.

The AlGa alloy also could not be exposed to the atmosphere without reacting with water vapour and producing very dangerous hydrogen gas. It's actually less safe that cryogenic or pressurized hydrogen because the reaction would be slow (and hence stoichiometric with atmospheric oxygen), and the reaction is exothermic. If a fuel cell car suffers a hydrogen tank breach at least the hydrogen will be high in the sky quickly. In this case of the AlGa powered car, the wreck wouldn't be safe to approach for a very long time.

Overall I think this is a terrible idea without any redeeming qualities and I'm very surprised by the limited amount of traction it received. Apparently the US Department of Energy has declined to fund this research, and evidently for good reason. If this wasn't a professor at Purdue, no one would have given this concept a second (or even first) look.


Fat Man said...

IANAC, but I think you need to reverse the sign on the second equation, but it still a bad deal.

Robert McLeod said...

No, those are all correct.

Anonymous said...

Shouldn't this be

4 Al (bulk) + 3 H2O → 2 Al2O3 + 3 H2

jeandaniel217 said...

I think there is actually a mistake.
In my opinion:
exothermical :
2 Al + 3 / 2 O2 -> Al2O3 + 1672 [kJ / mol (Al2O3)] (exothermic)
endothermical :
H2O -> H2 + O -285.8 [kJ / mol (H2)]
=> (3 for moles): 3 H2O -> 3 H2 + 3 / 2 O2 -857.4 [kJ]
Finally: 2 Al + 3 H2O -> Al2O3 + 3 H2 + 1672-857.4 = 817.6 [kJ]
The process efficiency is about 50% (857.4/1672)
But the Al electrolysis has also 50% efficiency too => about 25% for the complete process, without the fuel cell or the thermical motor.
A second point is that the electrolysis burn about 415 kg of Carbon for each 1'000 kg of Al (electrodes)!
That's the real bad deal ...
PS Excuse me for my bad english.

Anonymous said...

All false,

First of all, the water split reaction does not create alumina but AL(OH)3. And efficiency of aluminium electrolysis is close to 94% these days.

Anonymous said...

Your a little out of date the latest mix is down to around 20% gal with a small amount of indium and tin. The Aluminium in the new mix it closer to 75%