Embedded wafer level ball grid array (eWLB) technology has been a hot topic lately on 3D InCites, what with our recent discussion with ST Micro’s Xavier Baraton, and this week’s announcement by STATS ChipPAC that it’s ready to roll with 300mm manufacturing. We were lucky to catch the company’s CTO, Dr. Han Byung Joon, on his way back to Singapore, and got more of the details around the 300mm story, as well as what to expect for 3D eWLB.
STATS ChipPAC is developing next generation eWLB with technology partners Infineon Technologies AG and STMicroelectronics in an exclusive alliance that gives all three companies the rights to manufacture eWLB technology-based devices for their customers. Han explained that this minimizes the risk to each partner if more than one manufacturer is in production. However he said that to date, only Infineon and STATS ChipPAC among the three partners are in production with eWLB. ST Micro hasn’t invested in capacity yet, and hasn’t determined whether to manufacture internally.
So as it stands, STATS ChipPAC’s big news is that it’s launching its 300mm manufacturing line, and is the first in the industry. Infineon is running at smaller capacity on a 200mm manufacturing line. Han said the equipment making up the 300mm line comes from a variety of suppliers. To date the company has only announced a multi-tool purchase from Rudolph Technologies for eWLB inspection.
Developments in 3D eWLB
Our discussion turned to the next-generation eWLB, which involves taking the technology into the z-direction. This development work is also being done in Singapore. Baraton explained that essentially, they are creating a high-density Package-on-Package (PoP) by creating metallization on both sides of the package.
Han expanded on this, explaining that the primary benefit when compared with traditional PoP interconnected with wirebond or flip chip is the total package height. Currently the lowest height achievable with conventional processes is 1.4mm. “Each single eWLB package can be as low as 0.5mm, so we can come up with a product that is less than 1.0mm.” said Han.
And while Han thinks development of 3D eWLB will progress rapidly, he’s not sure market demand will happen as quickly as people think. At the moment, even high-end products like the iPhone can survive with the 1.4mm PoP. He predicts demand for packages below 1.0mm won’t increase until the later part of next year.
TSV and eWLB
How will the adoption of TSVs affect eWLB? According to Han, this could be a good thing, as TSV and eWLB are not competing technologies but could in fact be beneficial to one another. “TSV offers something that eWLB doesn’t.” he explained. “eWLB is a packaging scheme, while TSV provides another means of creating the 3D interconnections.” It’s a matter of manufacturing economics. The cost of silicon is shrinking, while the cost of package manufacturing is not. Economies of scale for eWLB are better than conventional packaging schemes. Additionally, TSV opens up new ways of making eWLB 3D, potentially increasing density in the same area, and that if TSV becomes a mature technology, it can be incorporated into eWLB. However, the full potential and cost of this concept is yet to be explored.
