It’s always a challenge to adequately sum up several hours of detailed discussion on a topic, and this afternoon’s symposium was no different. I’ve been told that I tend to give more detail than desired when recapping events. So this time, I’m going to attempt to break character, and take a less-is-more approach.
Today, it was all about 3D integration (not to be confused with 3D packaging, noted main liner, Dr. Phil Garrou), TSV processing, and reliability. Opening with a comprehensive, state-of-the-market status report, Garrou offered analysis and perspective worthy of the seasoned industry expert he is. Notable points to ponder from his presentation:
3D integration is an evolutionary process that began with face to face wafer bonding, no TSVs; followed by TSV technologies; now working on 3D stacking, because “It’s always easier to institute one new process in your foundry at a time than introduce several.”
Electrical performance issues around traditional scaling has lead us to where we are today; relying on 3D integration to further size reduction, enhanced performance, etc…
Memory stacks are next in line for 3D integration because memory manufacturers (Samsung, Micron, Elpida, Hynix) “all have to be looking in this direction or they’ll be losing significant market share in the future.”
TSMC claims it will be production ready with a 300mm TSV line by 2011 – Garrou thinks that’s pretty aggressive, and predicts in reality it will be 2012 unless you’re a “preferred customer.”
Inroads are being made in design….not much info yet on test, but in both cases, heads are up and taking notice. Garrou notes that in 2007, the design community’s response was ‘we’ll get involved when our customers indicate a need.’ In 2009, Mentor Graphics, Cadence, and Synopsis “have all bought in and are developing tools with ‘major customers.”
The afternoon continued with discussions of recent work being done at the U. Of Texas, Austin to determine thermo-mechanical reliability of TSVs, and an update on TSV technology implementation from industry pioneer, ALLVIA, and wrapped up with some progressive thoughts on the benefits of high-aspect ratio (HAR) TSVs, from Claudio Truzzi, of Alchimer.
Truzzi urged us to shift our frame of reference, and consider a different set of assumptions, based on design advantages of HAR TSVs rather than manufacturing limitations.“I believe 3D IC is a revolution, and cannot be achieved using evolutionary processes, as has been the trend,” he noted.
There’s a lot more to say about this, but it’s late and I’m tired. Think of this as the starter, with more courses to follow.