Sajid,
I would like to add a few comments and general information to supplement Eric
Sanjuan's suggestions on the plating setup. My suggestions are more solution
chemistry based, assuming an applied current density on the order of 50 mA/cm2.

In attempting a bottom fill or a superfill, the effective solution composition
at the surface is critical.
This is the reason why suppliers like Shipley and Enthone consider their
additives to be proprietary. In your case, the solution composition may not be
as critical, due to the relatively large feature sizes, but is stil important.

In a normal deposition, the current distribution concentrates at the edges and
on the surface preferentially, as opposed to areas of higher resistance (mass
transport resistance) e.g. at the bottom of a via. This results in a void
formation in the copper deposit.

The reasoning of using additives in the bath is to control the rate at which
copper deposits on the edges of the via and the top surface (suppress it), while
accelerating the rate at the bottom of the via. The commonly used additives are
as follows

Poly alkylene Glycol (such as PEG - Poly etheylene glycol). This acts as a
suppressant. How this works - These PEG molecules are large (Molecular weights
of about 1000). They are severely mass transport limited. As a result they like
to adsorb on the surface preferentially as opposed to the bottom of the via. A
typical concentration is 200 ppm and can range between 100-500.

Bis (3- sulfopropyl) Di sodium sulfonate commonly called SPS - Is considered an
accelerator in common terminology. But its actual role is that of a depolarizer,
meaning that it shifts the polarization curve to more negative potentials. In
the presence of PEG - it adsorbs preferentially over the PEG at the bottom (it
is a much smaller molecule and can hence diffuse much faster) The bi-sulfide
bonds cleave into monosulfides and promotes the adsorption of copper, by forming
an intermediate. The rate is lower than just plain old copper reaction, but is
much faster than that in the rate of reaction in the presence of PEG. In case
you wonder why the replenishments have to be done regularly to the bath -  its
because this sulfide compound is regularly adsorbed into the deposit. A typical
concentration is 50 ppm and can range between 40-200 ppm.

Chloride Ion- is known to improve surface quality. Why? Explanations vary. But,
Cl- is known to stabilize the Cu ion in a chloride electrolyte. It is also a
mild suppressing additive. Common concentrations are 50-100 ppm.


Other aspects of the bottom fill - Increase the ratio of Copper to Sulfuric
acid, by increasing Cu and dropping the sulfuric. (a) This reduces the
conductivity of solution - which is usually undesirable in an unrecessed, normal
deposition. But is desirable in a bottom fill approach, because (b) It reduces
the mass transport limitation of Cu, by increasing its concentration at the
bottom of the via. Solution pH of 2-3 would be better for your purpose instead
of the close to zero you have now.


Another aspect I considered was the non-continuity along the sidewalls, but 5000
A of seed should be sufficient to provide continuity along the side walls.

Of course the other important features of the plating set up are the current
densities and pulse parameters applied. These may be ascertained experimentally.
An on-reverse ratio of 5:1 for any current is a good starting point

Good luck and keep us posted on the results
Amrit.




-----Original Message-----
From: Eric Sanjuan [mailto:[email protected]]
Sent: Thursday, September 11, 2003 2:30 PM
To: [email protected]
Subject: [mems-talk] Re: Cu electroplating; sorry if you received this
mail twice (sajid)


hi Sajid,
            Just some thoughts.....
    The solution seems to be a high throw mixture (I think that's good for
your geometries) but your Cl- of 0.166mg/L seems very low, usually its in
the 50-80ppm range (=50-80mg/L). Some companies sell different additive and
solution specs. for DC, pulse or pulse reverse, make sure you have spoken to
them about your application (the Cl- concentration might be correct for this
additive system and/or maybe the additive system might not be ideal for
pulse reverse plating). Your pulse setting sound off too, they are too
similar. Usually its a long FORW. ~20ms and a short higher current density
REV ~1ms  (there's a lot more to this but this seems to be the basics and in
no way is 20ms and 1ms magic numbers).
            Some useful thoughts...... :)
1)Google search "pulse reverse copper plating"
2)Have you heard of the DOE software called JUMP? Do a small DOE varying
FORW., OFF, REV times and FORW, REV. current density and let the data lead
you to optimization.
Hope this helps. good luck,
                                                 eric


> Message: 7
> Date: Sun, 7 Sep 2003 21:00:46 -0500
> From: "sajid" 
> Subject: [mems-talk] Cu electroplating; sorry if you received this
> mail twice

> Hi,
>
> I need some help with Copper electroplating.
>
> I have 30 micron wide and 60 micron deep trenches in silicon and I need
> to electroplate copper into the whole trench. However, after a certain
> duration of plating, my trench gets closed off and I get voids in them.
> The best I've been able to get is a ratio of 2 in the thickness
> uniformity. That is 12 microns on each sidewall and 6 microns at the
> bottom.
>
> I do not use any mask for now. The copper is plated onto the whole
> wafer.
> I have tried many plating schemes such as pulse plating and reverse
> pulse plating with pulse periods of 12 to 13 ms.
> I have not tried DC plating however, because I thought it would not
> work.
> I have a seed layer that is deposited by sputtering and my estimate is
> that it is 5000 A thick so I believe the seed layer is not very
> resistive.
> I am using organic additives called PCM+ from the company LeaRonal.
>
> The plating bath has a magnetic stirrer.
>
> The solution is mixed as:
> CuSO4: 75 g/L
> H2SO4: 184 g/L
> PCM: .005 L/L
> HCl: .166 ml/L
>
> I chanced upon a solution to an electroplating problem by Eric Sanjuan
> on this website, and I'm hoping someone can help me.
> Also, how often would one need to make new baths.I mean in terms of
> Thanking you,
> Sajid
>