Geng:

You may be seeing the influence of masking of the bulk Mo from surface
contamination (or possibly the effect of a non-uniform oxide).  The reactive
SF6 could be highly selective to Mo relative to any surface contaminants
that might be present.  A difference in etch rate between the surface
contaminants and the Mo could lead to the island formation that you are
seeing.

You might consider an argon preclean step to remove surface contaminants and
to expose the bulk Mo prior to introducing the SF6.  Duration of a
required Ar preclean step will depend on the bias power level but is
typically 5-15 seconds.  You might try a long step initially to determine
whether the addition of this step influences your result but not so
long that you will sputter too much of the Mo.  You'll want to remove as
much of the bulk Mo as possible with the SF6, not the Ar.

If you do not have argon, oxygen might help if the surface contaminants are
hydrocarbons.  However, if the surface oxidization is the root cause of the
island fomation, the oxygen could pose a problem.  Nitrogen might also
work.  You are looking for a gas that will etch any surface contaminants
faster or at the same rate as the bulk Mo.

Robert

On Sun, May 1, 2011 at 5:07 PM, zhanggeng  wrote:

> Hi all,
>
> We are trying to dry-etch a finger electrode array (gap ~ 10um) with
> molybdenum(thichkness:~100nm). SF6 is used as the process gas under the RF
> power of 300W. The overwhelming majority of the unmasked Mo film could be
> "wiped off"(by reacted) well. But, several spots of Mo
> residual(diameter:~10um, even larger) hinder the completion of the task,
> leading to short-circuit ends. Subsequent wet etching with acid-mixture of
> HNO3 and H3PO4 was tried to deal with the problem but it could not work as
> wish. Can anyone suggest a better alternative or suggestions to deal with
> the Mo residual.
>
> Thanks a lot and best regards.
>
> Geng Zhang