Low viscosity potting compound

[Theo Markettos]

For various complicated reasons I have a chip in a ceramic PGA package with
direct access to the die and its bond wires. I wish to pot them with
something so that I can put moderate force on the bond wires without
snapping them (at the moment just touching a bond wire is enough to break it
from the pad at the package end).

So I think I need some resin to pot it. But because it's very difficult to
get potting compound between the bond wires which are probably 200um apart I
think I need some very low viscosity potting compound. I'd imagine trying
to put water on the bond wires wouldn't get through due to surface tension,
so we're probably talking something less viscous than water, or with a lower
surface energy (if I'm right in remembering this is the cause of surface
tension). Basically all I can do is introduce a drop and let it flow by
capillary action.

The other thing is that I'm worried that shrinkage during curing will be
each force to break the wires. Having something that is conductive is also
not an option, as I want the chip to work afterwards!

I think it's epoxy I need, but don't really understand the vicosity and
equivalent weight figures given the the manufacturers. Alternatively are
there other compounds that would do the same job (cyanoacrylate?). Whilst
I'm UK based, I'm mainly after information that would tell be where to start
looking. I only want a drop, not a gallon!

Thanks,
Theo

 

[Uncle Al]

For various complicated reasons I have a chip in a ceramic PGA package with
direct access to the die and its bond wires. I wish to pot them with
something so that I can put moderate force on the bond wires without
snapping them (at the moment just touching a bond wire is enough to break it
from the pad at the package end).

So I think I need some resin to pot it. But because it's very difficult to
get potting compound between the bond wires which are probably 200um apart I
think I need some very low viscosity potting compound. I'd imagine trying
to put water on the bond wires wouldn't get through due to surface tension,
so we're probably talking something less viscous than water, or with a lower
surface energy (if I'm right in remembering this is the cause of surface
tension). Basically all I can do is introduce a drop and let it flow by
capillary action.

The other thing is that I'm worried that shrinkage during curing will be
each force to break the wires. Having something that is conductive is also
not an option, as I want the chip to work afterwards!

I think it's epoxy I need, but don't really understand the vicosity and
equivalent weight figures given the the manufacturers. Alternatively are
there other compounds that would do the same job (cyanoacrylate?). Whilst
I'm UK based, I'm mainly after information that would tell be where to start
looking. I only want a drop, not a gallon!

Hasn't this question been addressed in your industry already? A month
in the lab will save you a day in the library. Beware ionic additives
and impurities! Free halide is the universal corrosive.

http://www.mpsupplies.com/3madhesivechart.html

Look up light- and UV-cured adhesives, especially optical bonders for
lenses. They are literally used by the drop and 100% cure within
seconds but have almost unlimited pot life in the dark. Shrinkage can
be reduced by adding micronized filler, but that can boost the
viscosity.

There are a number of resins whose cure is based upon opening of
spyrocyclic monomers (that expand on curing) instead of linear
concatentation of small monomers (that shrink on curing). You can
thus get essentally zero volume change - for a price - in several cure
system mechanisms. Beware ionic additives and impurities!

One cute possiblity is to take a low viscosity epoxy, urethane,
photocure, or whatver and add a modest loading of dense micron-radius
glass beads used to make reflective coatings. You pot your device
with well-mixed filled resin and let the potting settle, maybe with a
thump or three to close-pack the beads at the bottom by the chip.
Cure. Now you have a physically robust potting with little local
shrinkage, the whole shrink-wrapped by overlain unfilled epoxy.

Google
adhesives -epoxy cure 31,500 hits

 

[Mark Thorson]

One cute possiblity is to take a low viscosity epoxy, urethane,
photocure, or whatver and add a modest loading of dense
micron-radius glass beads used to make reflective coatings.

That won't work. Glass will contaminate the encapsulant
with sodium, which eats up aluminum bond pads.

Best solution is to pretend to be working for a semiconductor
manufacturer, and request samples from encapsulant suppliers.
Pure epoxy would be nice for its low viscosity, but it's
usually supplied with lots of silica filler and formulated for
injection molding (an exception would be Thermoset, who
offers semiconductor-grade pure epoxies). Silicones tend
to be higher viscosity, but are widely available in semiconductor
grade (Dow-Corning, Wacker, Shin-Etsu). Polyurethanes
are available, but combine the worst features of both silicones
and epoxies.

What you are asking for is commonly called a "glob top"
encapsulant.

Before selecting a UV-cure system, make sure your chip
can tolerate it. Flash memory is permanently disabled by
UV light (it erases critical data held in flash memory cells
outside the memory array, such as read threshold voltage).

Glob tops are generally filled with carbon black to make
them opaque. Is that a problem?

 

[jacques jedwab]

Experiment with plain parafine, which fuses at low T.

J.J.

 

[Uncle Al]

Experiment with plain parafine, which fuses at low T.

Paraffin wax has an amazing shrinkage upon crystallization. It
continues to ripen over time.

 

[Greysky]

What about a drop of 'superglue'? Then you can apply some other material
over it to complete your seal.

 

[Mark Thorson]

What about a drop of 'superglue'? Then you can apply some
other material over it to complete your seal.

Not a chance, unless you only need the chip to survive for
a few hours or days. The aluminum bond pads are sputtered
aluminum, and they are so thin they will be eaten up by even
low levels of ionics. They are not stable in the presence of
(at most) 70 ppm or more of ionics (Na+, K+, NH4+, Cl-, etc.).

Very few materials that are not made for contact with a
silicon chip are this ionically clean. Some silicones may be
this clean, but other than that, forget it. And if you use
silicone, use the platinum-catalyzed addition cure type,
to avoid generating reactive species.