In this link you can find a complete and clear review about aluminium anodizing:

http://www.wiley-vch.de/books/sample/3527318763_c01.pdf

Thank you. Do you have any literature on barrier-type anodisation of Al?

In this work, the first part is about barrier-type films:

http://www.sciencedirect.com/science/article/pii/S0040609096094400

Hi Leszek,

Aluminum anodizing uses fairly aggressive wet chemistry, and generally is done on complete surfaces.  regions can be masked so that they do not anodize, again usually on a large area sort of scale. I think this technology would be challenging to implement at transistor like scales.

   The problem that you will encounter, which will be much worse at the small scale is auto-oxidation of aluminum.  no fresh surface of aluminum will remain a pure conductor for very long. The accumulation of Aluminum oxide at the surface will convert it over time from a conductor to a semi-conductor to an insulator.  The rate of this change is driven by a number of external factors.   As a result of this oxidation , Aluminum is also difficult to make a solder connection to.

In short there are reasons that aluminum is typically not used in integrated circuit manufacturing.

Regards,

Brian

Aluminum has been used extensively as interconnect in the past, only the need to drive higher currents in smaller packages has gotten us to copper. Copper is actually bad, because it will diffuse into the Si layers and start doping.

So, while theoretically possible, there are several things that speak against it:

a) If you can avoid doing wet chemistry, you do. It's messy, and you need to clean your mess up afterwards.

b) SiO2 is beautiful because: thermal growth and easy etch with HF (even HF vapor). Al2O3 is very difficult to etch, so you have to Change all your processes around. Front End of Line (FEOL) is allergic to that.

c) Better than wet Deposition is sputtering. But then you take Ta2O5 with double the dielectric constant.

In summary - the advantage for organic electronics of doing everything wet is not really an advantage in the world of silicon, where you prefer to do everything dry, especially in the FEOL. Also, Al is usally deposited in the BEOL as an interconnect (when not using Cu), and they are very allergic to metal deposition in the FEOL due to possible doping (remember, Al makes a nice n-dopant for Si, although it does not diffuse as much as Cu). And if you are already sputtering, use Ta2O3...

My thoughts are that in MEMs Technologies, with larger scales, Al2O3 may be a lot more viable...