John,
Another solution to your problem is to use a heat-spreader with a better CTE
match to silicon and the substrate. Since your substrate is alumina, and power
dissipation is high, the CTE of the heat-spreader should be somewhat higher
than that of Alumina, but not as high as the CTE of Aluminum. There are ways
of determining the optimum CTE for a given substrate material. WCu and AlSiC
heat-spreaders are CTE-tailored (with a cost premium) for that very purpose.

Bottom line is that your adhesive failure problems will go away if the entire
structure (substrate/die/ heat-spreader) is "CTE-balanced".

With best regards,
Jean-Paul
_____________________________________________________
Jean-Paul Clech
EPSI Inc., P. O. Box 1522, Montclair, NJ 07042, USA
SMT / Packaging Reliability Consultants
tel: +1 (973)746-3796, fax: +1 (973)655-0815
home page: http://members.aol.com/Epsiinc1/index.html


In a message dated 98-11-19 05:31:10 EST, you write:

<< My question relates specifically to a heatsink bonded
 to the rear surface of the silicon die on top of a Motorola Power PC (which
 uses IBM C4 technology) on a ceramic BGA substrate.  The silicon surface
 area available for heatsink attach is approximately 10mm square and the
 chip power dissipation is 7W.

 For previous generation (non BGA) products we used adhesive thermally
 conductive tape bonded to the top of the ceramic pasckage, which worked
 fine as the surface area was about 30mm square. With the BGA package
 structure the bondable area is down to 10mm square but the heatsink has to
 stay big (20mm square) to get the required heat dissipation.

 Motorola sort of recommend mechanically clipped heatsinks, but these are
 not practical in our application due to constraints of size and
 availability, leaving us for at least the short term with the need to find
 an adhesive that can cope with the mismatch in CTE.  Tape does not give
 enough bond strength.

 We are doing trials at present with Berqquist LA1000, a thermaly conductive
 low modulus adhesive, which gives good bond strength.  The manufacturer
 offers versions with glass beads (4.9mil or 7mil) for control of bond line
 thickness, which may or may not be useful.

 We would like to know more about the long term reliability aspects of this
 adhesive. Does anyone have any data in similar applications?
  >>

################################################################
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
################################################################