It’s not often that I get to meet other members of the semiconductor aristocracy, so I was delighted to recently find myself in the company of Datacon’s Hugo Pristauz, the King of Flip Chip. It was a mutually advantageous meeting – I introduced him to the 3D InCites community, and he offered some valuable insight on what it’s like to successfully propel an emerging technology from R&D to volume production. Although this isn’t the first time I’d heard the progress of 3D TSV commercialization be compared to flip chip’s journey, it was the first time I’d heard it from the perspective of an equipment supplier whose tools are targeted for high volume manufacturing, without a toehold in R&D.
I am not a stranger to Datacon’s flip chip success story. While visiting facilities as part the Advanced Packaging Roadshow, it seemed every applications lab or full assembly line I toured boasted a Datacon die bonder or flip chip bonder. But Pristauz explained that it’s only in the past few years that Datacon has fully realized the benefits of flip chip becoming mainstream, receiving tool orders in production-level quantities. And although it seems to be a natural segue for the company to go from the flip chip market to the 3D TSV market, the current R&D activity that provides a revenue stream for companies like SET, EV Group and SUSS MicroTec, who have tools either target the R&D environment, or are designed to evolve from R&D to a volume production environment, doesn’t really serve Datacon. While Pristauz assured me that 3D TSVs are firmly on Datacon’s roadmap, he also remains cautious about its commercialization, echoing Steve Lerner’s sentiments about a need for reduction in TSV cost of ownership before that can happen.
That’s not to say Datacon hasn’t been active in developing pick-and-place solutions for 3D stacks. They are members of the EMC3D consortium, and for the past few years have worked in partnership with EVG to develop a two-step chip-to-wafer process that will help solve the speed/accuracy limitations. They’re working on the pick-and-place accuracy while EVG works on the gang bonding step. Additionally, last year we reported on the company’s collaboration with TNO for its Bluebird Project on a similar project with the ultimate goal of developing a high-accuracy, production level flip chip bonder with a placement accuracy of 2.5µm at 3sigma. According to Datacon’s Hannes Kostner, who participated in a discussion on the C2W progress a few weeks ago, the company has succeeded in increasing accuracy from 10µm at 3 sigma to 7µm at 3sigma on 250mm wafers, and are working towards 300mm.
Additionally, in other areas of 3D WLP, Pristauz says Datacon’s flip chip bonders dominate the market for eWLB, with tools in place at all the major eWLB manufacturers. As eWLB moves towards its next generation in 3D, this is a good position for Datacon to be in.
I suspect we’ll start hearing more news about Datacon in relation to 3D integration. Who knows, maybe Pristauz’ flip chip kingdom will expand into the 3D realm as well, although I’m not ready to abdicate that throne just yet. — F.v.T.
