Ingemar,
few things you may want to know:
(1) your wire bonding yield may be going down initially (your manager will
scream down to your neck, it is normal process)! This is related to the
requirement of modified bonding parameters for the soft gold which you may
obtained after the optimization of your plating process. Get your best
operator work on the experimental setting... if he tells you, the gold is
"grab" the wire bond rather than "slip" on the surface, you are almost there...
(2) for 1.25 mil Al wire to gold pad, the acceptable lower pull strengh is
9g and upper limit is 15 g with nice peak average around 12g (nothing less
than 9g). This is the first sight of "luck is smiling at you". However,
(3) do not celebrate the event unless your Au bath is 3 weeks old, you can
still obtain the 9-15 g result. Everybody (ya, ya, may be not everybody)
can plating nice Au pad when they are using brand new bath. Stablization
and control is the key. (you may have a glass of wine and take a nap, you
may still have nightmare about wire bond).
(4) When you establish nice SPC chart and everything fall into 9-15 g (for
1.25 mil for example), your manager will stop make notice of you...(good or
bad? but anyway, take a break). You may want to thermal age your wire bond
and make sure the minimum strength is 6 g, nothing less...(minimize the
kirkdendahl void, etc. you may want to check something normally overlooked
by many, such as grainboundary diffusion, grain size distribution, surface
ionic contamination etc. etc.). When you get this far, you may pack
yourself on the back and open a bottle of Burgundy. If I were you, I would
have a long sweet dream and try to forget everything about wire bond....
Good luck and see you at end of the fiber 6-9 month (if you can hang in there).
jk
At 12:16 PM 5/5/99 +0200, you wrote:
>>HI JK,
>felt that I had to tell you that I'm using your comments. Yes, I'm
>sucking and tasting and thinking. Maybe I will be more wise than before.
>If something more appears, I will let you know. I use to come back, say
>thank you and tell what happened next. But it can take some time. See
>you at the end of the fibers./Ingemar
>
>
> Dave hit the nail on the head. The difficulty of wire bond on gold may
>have
>> many root causes. Here are my 2 cents:
>>
>> (1) Plating problem. Plating condition, e.g.current density, PH,
>additive
>> concentration variation will resulting entrapment of impurities
>> (co-deposite), sometime the Au deposite become porous. A good
>reference
>> paper for Au plating is:
>> High-Performance Gold Plating for Microdevices - by A. Gemmler
>> et.al. Plating and Surface Finishing, Aug. 1994, p.52
>> If your vendor perform the similar analysis as indicated in the paper,
>you
>> would not have any worry about the wire bonding for a long time (You
>will
>> notice the change on the "wire bonding yield chart" if someone
>changed job
>> on your vendor side. for example, an engineer or QA...or someone
>decided to
>> go "out sourcing", cut cost to "extend Au bath life", etc...No
>kidding!)
>>
>> (2) Surface contamination and surface diffusion. Poor coverage at the
>edge
>> of the gold pads (exposed Cu or Ni) sometime will result Cu/Ni
>migrated to
>> Au surface under humidity condition. (surface diffusion is 2 order of
>> magnitude high than the bulk diffusion).
>> Insufficient final rinse (or delay of the final rinse, e.g. "went for
>> lunch") after the plating may cause residue on the gold pad (oxidation
>of
>> residue...e.g.brownish look Au for KCN based plating). Good
>reference:
>> The Auger Analysis of Contaminants that Influence the
>> Thermocompression Bonding of Gold. - by G.E. McGuire et.al, Thin Solid
>Films
>> Vol. 45, 1977. p. 59
>>
>> (3) Plasma cleaning does have some draw backs: You will activated PWB
>> surface too. Plasma clean is a standard cleaning method used for
>ceramic
>> thin/thick film product prior to wirebond. The work time after the
>clean is
>> limited to few hours (depend upon what class of clean room you are
>using).
>> It is costly to do surface analysis (Auger, SIMS, XPS) to define the
>working
>> time after plasma clean. I usually specify on the low side (ASAP=less
>than
>> 0.5 hr). UV-Ozone may be faster, but PWB damage have to be assessed
>if you
>> are using FR4 or polyimide.
>>
>> (4) Hardness of Au is a quick and cheap way to define high quality Au
>(it
>> does not tell you much about surface contaminant). However, it is not
>easy
>> to do micro hardness (you need special machine! $$$!). Particularly,
>when
>> thickness is on the low side (< 30 micro-inches).
>>
>> (5) It is difficult to do cost-benefit assessment for plasma
>equipment.
>> Quality improvement normally take about half a year to show some
>measurable
>> result. I face the same issue too. Please accept my condonlence.
>>
>> my 2 cents...Good luck.
>> a pupil (jk)
>>
>> p.s. regarding Au quality, you may want take a look of "The use of
>SIMS to
>> investigate wire bonding yield problem on gold contact" by G.R. Mount,
>> et.al. MRS Symp. Proc. Vol. 390 Electronic Packaging Material Science
>VIII,
>> 1995 p.245
>> ---------------------------------------------------------------------
>> At 07:55 AM 4/21/99 -0700, you wrote:
>> >All this back and forth conversation is great regarding how to bond
>to Au.
>> >Remember the first rule of thumb. GET TO THE ROOT CAUSE!!!!
>> >
>> >A eraser is simply abrading the surface, removing "crud" but also
>removing
>> >precious Au that you need for acceptable bonds. The eraser does not
>> >miraculously act as a vacuum but a tool to smear the crud, not remove
>it.
>> >If it is an organic problem that you have, plasma may work but plasma
>will
>> >only remove ~200-300 angstroms, not a whole lot being removed.
>Plasma also
>> >depends on the gas, flow, temp, electrode configuration, chamber
>size. You
>> >just can't buy a plasma unit and plug it in.
>> >
>> >Good Luck and if more help is needed, I can be reached at
>> >[log in to unmask]
>
>> >---------------------------------------------------------------------
>> > Hi, sorry to get in between. You seem to discuss what is hot also
>here
>> >at our company to and from. My first question would be: please,
>> >gentlemen, what are you from Chuck and pupil? I have got some good
>> >advice from Dave at Rockwell, maybe we have something good to learn
>> >from you too. I have tried for one year to buy an american plasma
>> >cleaner, but when the economists ask "what is the payback time" ,
>then
>> >what can I answer? Oh, my adrenaline...! /Regards
>> >
>> > Ingemar Hernefjord
>> > Ericsson Microwave Systems
>
>
>>
>
>################################################################
>TechNet E-Mail Forum 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's web site (http://www.ipc.org) "On-Line Services" section
for additional information.
>For technical support contact Hugo Scaramuzza at [log in to unmask] or
847-509-9700 ext.312
>################################################################
>
################################################################
TechNet E-Mail Forum 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's web site (http://www.ipc.org) "On-Line Services" section for additional information.
For technical support contact Hugo Scaramuzza at [log in to unmask] or 847-509-9700 ext.312
################################################################
|
