Hi Alessandro Cavalli I am assuming the floated CuInSe layer is the important part of your work? I am not sure why you are using Al2O3? One of the features of Al is that it forms oxides very quickly in atmosphere that form tough protective layers. In "Advanced Inorganic Chemistry" p 262, 1972, Coton and Wilkinson, it states that Al will dissolve in H2O. So I am wondering if Al would be a better lift-off layer? I successfully modified my own lift-off mask process to accommodate this property of Al.

I'm afraid the choice of Al2O3 as a 'sacrificial layer' was one of the worst options ever. A proven way of removing Al2O3 is a plasma etching in BCl3/Cl2 mixture but it should severely damage your copper-contaning layer as well.

As Martin suggested, Al interlayer would work nicely with a wet etching - though not in a pure water but in some alkaline solution , e.g. photoresist developer. Additionally you should check that this recipe wouldn't damage your material of interest.

Most probably you're restricted in the material selection because of the subsequent MBE process which might not tolerate any metal layers. In this case deposited in either way SiOx or SiNx might help.

Victor V Solovyev Yes, the use of NaOH would work. What I forgot to mention before is that if you etch Al and remove the surface oxide and keep the sample wet you might then be able to remove the remaining Al in water. In my work I was etching Cu under Al and found that a diluted Cu wet etch exposed bare (non-oxidised) Al which then dissolved in the Cu etch. A more concentrated Cu etch helped address this.

I do not know much about the chemistry of CuInSe , so my proposal may not work (i.e. would possibly destroy your precious CuInSe layer), but Al2O3 [when not having been grown in a high-temperature process (and e.g. rather by evaporation, sputtering or ALD)] can normally be etched extremely well in hydroflouric acid (HF).

Hi Ale, not sure if still required, but similar to Nils, I'm not sure if CuInSe will be attacked. But for me, 2% BOE ( buffered oxide etch~ HF) had a slow enough etching rate (0.1nm/s) for ALD grown Al2O3 to get well controlled stable etch and leaves the Si wafer intact. H3PO4 or AZ400K were too agressive. Hf will be a pain to work with though, but 2% should be ok to handle if you can dilute later with sufficient H2O...

Hi everyone,

Thank you very much for your answers.

Martin Goodchild and Victor V Solovyev : unfortunately Al2O3 is an important part of my process, and in particular SiN or SiOx do not work as well as Alumina. I have not tried to use Al, but indeed in MBE I will have some issues by growing in pure Al I think, especially since I use temperatures above 600C that will damage the Al.

Nils Harder and Julian Treu, thanks a lot for these suggestions. I am also not sure about the chemistry of CISe, but I will try and see what happens by using HF. Julian Treu can you share anyway some details about the H3PO4 etching? Maybe it is not too aggressive for my purposes.