I will contribute one thing to this complex issue:

One thing I know - At low current densities one achieves the best
uniformity, but not necessarily the best morphology or properties.

Regards,

David Roberts (Wafer Plating Specialist)
Prodigy Surface Tech (408) 469-3203 cell
807 Aldo Ave., Suite 103
Santa Clara, CA 95054
Phone (408) 492-9390
FAX (408) 492-9391

-----Original Message-----
From: Oakes Garrett [mailto:G.Oakes@EVGroup.com]
Sent: Monday, July 06, 2009 9:56 AM
To: General MEMS discussion
Subject: Re: [mems-talk] Metal electrodeposition rate

It has been my experience with copper electroplating (35 µm tall structures,
50 µm thick resist, predominantly circles with some crosses and short lines,
Cu/Ni seed) that features with CD's below 20 µm will plate at different
rates that features with CD's above 20 µm.

I studied a case in some detail (from a litho perspective) a few years ago
with a range of feature sizes (5 - 50 µm at 5 µm increments).  As stated
above, the 20 - 50 µm features plated at roughly the same rate across the
entire wafer surface.  The 15 µm features showed some decrease in plating
rate and the rate continued to drop with shrinking via dimension.  Diffusion
of copper ions to the plating surface was our suspect for the reduced
plating rate.  We didn't believe at the time that we had any wetting/trapped
air issues that would lead to a delayed start or stall in plating.

Jumping to a conclusion, the important value is most likely the aspect ratio
of thickness to lateral dimension.  This one case suggests that a ratio of
roughly 5:2 is the critical plating aspect ratio for circles.

My response is a very short treatment of a complex subject.  I would like to
see responses from other people with data to see if my facts check.

Best Regards,

Garrett Oakes