> Hi everybody,
>
>      Does it have any probelms if we deposite the metal(gold) on the =
> poly1=20
>  without depositing poly2 using MCNC MUMPS?=20
>
>     Recently, we had some MEMS dies back for testing. But if the design
>  deposited the metal on the poly1 without poly2 (i.e. poly1+via+gold =20
>  NOT poly1+via+poly2+gold) , the design seems to stick the wafer.=20
>  And after pushing  it with probe, the structure will fail..... Does =
> anyone=20
>  knows what the problem is ? Thank you!!!
> =20
>
>
> Sincerely,
> ----------------------------------
> Yu-Chen Lin
> Electrical and Control Engineering
> National Chiao-Tung University,Taiwan
> u8612519@cc.nctu.edu.tw


MUMPs does not support metal directly anything other than poly2; it will
not adhere.  You may also have problems with the depth of the liftoff mask
definitions, depending on your layout (either undercut to areas larger than
defined, or too thick for good lithography.  Also, beware of metal that is
bridging unsupported over an oxide: these can break off and get stuck
elsewhere.

I know you have another problem, here:  poly1+via+gold will etch away
poly1, due to an overetch of poly2.  This is designed into the MUMPs
process (and others) to prevent stringer formation.  If you've ever seen
pictures of our stacked-poly devices, this is how we form 2.5um thick poly
devices with a single mask.  A similar thing happens in MUMPs when you etch
p1p2via from the top of oxide1, some (but not all) of the first oxide is
etched away.


 - dave.

; Center for Advanced Computing and Communication,  NCSU Box 7914,  ECE-EGRC =
;-   __o   =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
;  _`\<,_   To build a computer that knows more about itself than I, won't   =
; (*)/ (*)  [(919)513-2015]   require much thinking,  --+--   ... I think.   =
; [http://www4.ncsu.edu/~dawinic]  `---================(-)================---'