Thanks for your kind interest, Carey.
jk may be glad when I say a thinfilm PhD is welcome. Carey, I agree, the interface between the members is where joining really begins, bulk material most oftenly being more stable and homogenous than the hazardous and caotic meeting between atoms. If they don't like each other there will be no connection, simple image but a theoretical mess. Carey, U/S wire bonding, as you know, is in practice small locations where all debris, films and other stuff has been removed so that metal binding can occur. These microwelds are only nanometer-to-micrometer large and the percentage of adhesion between ball and pad is amazingly poor, only some percents, the major areas only flattened and "machined" but with still no binding. Now, IF you realize that a maximum of these much valued microwelds are wanted, a flat surface (thinfilm gold) seems to be superiour to a rough surface (thickfilm gold), the reason being, not thinfilm intermezzos, but a question of the total availability of metal. So, !
!
!
if the gold pad is too "wavy", the bond capillary has a tough job to press metal against metal. If you have a rough surface to bond on, my thought was that having larger balls would increase the possibility to make more dates with the partner, literally spoken. Now, back to your adhesion theories, of course you are right, it might be a thinfilm problem in reality. My following question is therefore this. What kind of film is the worst troublemaker? Water, which always exist in some molecular layers thickness. Hydrocarbons, which also always exist everywhere. Oxides? Carbonates? Gases? Reactive metal atoms? Etc. And what method do you use for indication? Auger is always mentioned, but there are lots of other sophisticated equipment with names that can't even be pronounced. Our unis have most of them, but with 2 kilograms of diagrams in your hand. Your problem is still not solved, it's like getting a map but no compass when you are lost in the jungle.
If I send you a sample on our today's headache, can you then make the aspirine? I will send by DHL if you give me your adress, offline if that's more comfortable for you. All in the business have ideas, but I want to get to the point this week.
Thanks in advance
Ingemar Hernefjord
Ericsson Microwave Systems
Ingemar
Your problem is in basic metallurgy. The bond/bond strength failures are
not occurring in the bulk of the pad, but rather either at 1) the
interface of the Au-pad/substrate or 2) poor sonic-bonding parameters.
About the latter, thicker does not usually help for a variety of reasons.
One is that thinner films actually are stronger than thicker ones (although
25 um is nearly bulk quality depending on deposition methods). In
metallurgy speak and generallizations, there are more dislocations pinning
metal crystals into place in thin films. Another is the bonding process is
mixing the interface of the wire with the bond pad along the side of the
wire onto the pad. It's not like a shaft getting pushed down. So, deeper
is not better since you are effectively cratering out the side support that
you think you are getting. Also, your bonding parameters using the
"springiness" of the film to break the surface (if pure Au, then its to
break through the surface impurities that lodged into the surface).
Thicker wire means you have to recalibrate your
whole system (i.e., pad thickness) since you need to change the bond
strength that flattens the wire. High surface roughness (your Ra=5um is
20%... very high) gets in the way of
having a reproduceable process since one can't control uniformity.
I don't understand what you mean by Au/blue dielectric/Au. Are these
capacitors? The combination doesn't make sense from an IC point of view.
And what is the blue dielectric?
However, before embarking on this, you need to determine where the failure
is occuring. Dirty interfaces are usually the culprits. Burnishing makes
me real uncomfortable in this regard.
Carey
ps- thin films/interfaces are my expertise (Ph.D)
-----Original Message-----
From: Ingemar Hernefjord (EMW) <[log in to unmask]>
To: [log in to unmask] <[log in to unmask]>
Date: Friday, September 24, 1999 5:33 AM
Subject: [TN] Wire bonding question
>Is there a wire bond specialist listening?
>I am frustrated about a MCM package, bottom Al2O3, seal frame Kovar,
leadframe Kovar and conductors/mounting pads consisting of either a)one
layer Au thickfilm b) Au thickfilm/blue dielectric/Au thickfilm. Problem:
bondlift or bad bond yield. Automatic 25micron U/S ball bond K&S. The Ra is
from 1-5 microns. Conductor/bondpad thickness 15-20um. Surface roughness
eliminated somewhat by burnishing. Recommended (duPont) Ra-value is 0.5um
for good yield. Lots of Si, Bi and Cd particles but <10um large. Of course
these oxides and binders are not wanted, Au should be 100% pure, but as all
know that does not exist in practice for thickfilm. Now to my question, dear
Mr/Mrs Bond : if a gold thickfilm is rough, wouldn't an increase from 25
microns to 50 microns make the bond conditions a lot better? My simple
logic: rougher surface>thicker wire. With all admiration for Harman's
"Bible", there is nothing mentioned this way.
>Manipulating the force, amplitude and capillary face are all tried with
minor improvement, at least far from 6Sigma goal. Plasma cleaning or any
other cleaning does not help either.
>All answers are welcome, even from the small island (forgotten the name)
where Aussilek tries to live.
>Ingemar Hernefjord
>Ericsson Microwave Systems
>
>##############################################################
>TechNet Mail List provided as a free service by IPC using LISTSERV 1.8c
>##############################################################
>To subscribe/unsubscribe, send a message to [log in to unmask] with following
text in
>the body:
>To subscribe: SUBSCRIBE TECHNET <your full name>
>To unsubscribe: SIGNOFF TECHNET
>##############################################################
>Please visit IPC web site (http://www.ipc.org/html/forum.htm) for
additional
>information.
>If you need assistance - contact Gayatri Sardeshpande at [log in to unmask] or
>847-509-9700 ext.5365
>##############################################################
>
##############################################################
TechNet Mail List provided as a free service by IPC using LISTSERV 1.8c
##############################################################
To subscribe/unsubscribe, send a message to [log in to unmask] with following text in
the body:
To subscribe: SUBSCRIBE TECHNET <your full name>
To unsubscribe: SIGNOFF TECHNET
##############################################################
Please visit IPC web site (http://www.ipc.org/html/forum.htm) for additional
information.
If you need assistance - contact Gayatri Sardeshpande at [log in to unmask] or
847-509-9700 ext.5365
##############################################################
|
