Ever since my Advanced Packaging days, I’ve been drawn to follow what’s going on with the start-up companies. Their news is exciting. The people who work there share a certain enthusiasm that is contagious. It’s also a lot more fun to write about breakthrough technologies then improvements on legacy processes. That’s the main reason I chose to focus on 3D technologies in the first place. So at last week’s 3D Architectures for Systems Integration in Burlingame, CA, I sought out a few of the more mature start-ups in the 3D TSV space to see how this past year has been for them. Here’s what I came up with.
Mark Welsh of NEXX Systems was happy to report that the company had a banner year, culminating with its 100th system shipment, an electrodeposition tool now being installed at a leading Taiwan fab. Welsh notes that they didn’t have ONE single layoff in 2009, actually increased staff by 25% (who else can claim they did that that last year?) and outsourced work to local businesses (they’re located in Billerica, MA) stimulating the local economy.
2009 revenues for NEXX actually doubled from 2008, and 2010 projected at +30% over 2009 revenues. All this, achieved in a short 8 years of existence. Here’s a good example of a star-tup that hit its stride as many mature industry suppliers were forced to scale back. NEXX Systems was visionary about advanced packaging processes, and positioned itself well.
Tezzaron Semiconductors also experienced growth in the down turn, and at 10 years young, clearly has hit the same sweet spot as NEXX system. Bob Patti, CTO and founder, reports an increased revenue of 500% from the year before. They have also added staff, formed an alliance with R&D foundry SVTC in Austin TX for backside processing, and had success with its multi project wafer program (MPW). Tezzaron is known as an early adopter of TSVs, and have been pioneers in many areas.
Cataloging recent achievements, Patti talked about the progress 10 years of trial and error has brought about. He said problems using copper to fill via brought about a switch to tungsten. The company has perfected a process by which you can split layers. They’ve got a supply chain established and can build working, reliable parts. They’ve got a back-end tool flow and boast numerous early adopters: more than 65 customers in all. While everyone else complains about design and test issues, they’ve already integrated MicroMagic’s layout editor, and have designed their own built-in self test and repair (BiStar) technology that Patti says “is an innovative way to improve the yield of highly parallel structures such a memory” because when you’re dealing with 1.5M TSVs per layer, you “never touch the bond pad.”
Tezzaron’s roadmap for commercialization of 3D devices? CMOS image sensors are happening now. The near term will bring 2/4/8 layer memory followed by a logic/memory combo (putting their memory on other company’s 2D logic devices). In the next 2-3 years we’ll see mixed signal/logic for smart power, and in 5 years, look for logic/on logic.
Ziptronix is another lucky winner in the 3D landscape, especially with its recent coup of licensing its patented Direct Bond Interconnect (DBI®) 3D integration technology to Raytheon Vision Systems for manufacturing the company’s high-end image systems. Calling DBI an “ideal technology platform for 3D IC manufacturing, Paul Enquist, CTO and V.P. of R&D, Ziptronix, explained the technology as a “heterogeneous bond between two surfaces comprising a conducting component providing electrical interconnection, and an insulating component isolating electrical interconnections.”
“We believe we’re the only company that’s announced a manufacturing license for true 3D integration,’’ said Dan Donabedian, CEO, Ziptronix. “We are currently transferring our technology to enable Raytheon to go to production.” This is significant for two reasons, noted Donabedian. First, it demonstrates recognition of the IP itself, and secondly while recognizing the IP, the licensing company (Raytheon) needs to do their own due diligence.
Also reporting progress in the commercialization of TSVs was ALLVIA, with the purchase of a manufacturing facility in Hillsboro, OR for high-volume production of their products with TSV technology. The former Applied Materials facility is 178,000 sq. ft. with 60,000 sq. ft. (expandable to 80,000) of cleanroom capacity. This is 10X more space than the current facility in Sunnyvale. ALLVIA founder, Sergey Savastiouk (credit with coining the term “through silicon via”), saw a a need for quick turnaround prototyping of TSV wafers. Up until now a company’s choice is to either develop a facility of their own or go to the big foundries focusing on TSVs. There’s not enough experience in manufacturing wafers with TSVs. “Before people learn to run, they have to learn to walk,” is Savastiouk’s philosophy. ALLVIA can fill this need, and with the larger facility, will be able to handle high volume production.
Savastiouk believes silicon interposers will be the quickest to benefit from TSV technology, as they allow for die to be placed directly onto the silicon surface and then connected to the substrate, eliminating CTE mismatch. ALLVIA has developed an active silicon interposer device that incorporates a thin film capacitor rather than a discrete capacitor, thereby delivering 1000 nanofarads/cm2, notes Savastiouk. providing a solution to the problem of supplying power to high-speed chips.
All these companies have been in existence for no longer than 10 years, were established to support emerging 3D integration technologies, and have expanded in a year when most companies have had to cut back. Their progress and success should be an inspiration for some of the younger start-ups in this market. Kudos to all! – F.v.T