Morning all.

I'm trying to etch a thin (15nm) epilayer of MnAs to stop in a thicker
(300nm) layer of GaAs.  The problem is that while MnAs wet-etches
rapidly with almost anything (and the etch spreads laterally in an
uncontrollable way), it is amazingly resistant to dry etching (not just
my observations, from the literature too).

I've tried 'RIE' using just argon (should I call it 'IE', since I don't
want 'reactive'?), to create a physical etch and (hopefully) end up with
something that might be slow, but will at least be not particularly
selective between MnAs and GaAs.

It isn't.  I'm still getting maybe 20x selectivity between the two
(thanks to a very slow MnAs etch rate), despite the manganese atom being
slightly lighter than gallium (resist goes another order of magnitude
quicker, but I'd thought since resist has all those nice organic light
atoms, this isn't surprising).

Etch rates in all substrates are super-linear with power, which I
interpret naively as being more like a linear relation with accelerating
voltage.

Which (at last!) brings me to my questions.  What should I expect in
very general terms from a purely physical plasma etch?  Is it dominated
by chemical bonds in the substrate rather than atomic mass?  Might there
be a reactive contaminant in my system (I believe not)?  Is it even
plausible that my highly magnetic manganese interferes locally with the
plasma?  Better still, has anyone discovered the holy grail of MnAs dry
etching (reactive or otherwise)?  ;-)

Comments and corrections would be much appreciated.

Thank you!

Andy