July 1998
eMail the Editor
Mark DiOrio on Assembly: An Expert Looks at the Issues
Editor's Note: We interviewed Mark DiOrio for this special issue on IC package assembly. Mr. DiOrio is president of OlympIC Technologies Inc., a company providing a variety of advisory and support services to the semiconductor community in technology, management, strategic partnering, investments and offshore ventures. Prior to forming OlympIC Technologies, Mr. DiOrio occupied a variety of senior executive and management positions at ESEC, Disco, National Semiconductor and Olin. He holds degrees in mechanical engineering and materials engineering from the University of Connecticut. He is also a past chairman of SEMI's International Packaging Strategy and Roadmap Committee. Readers may contact him at 408.253.9972 or by e-mail at mark@olyic.com.Q What are the key challenges for the assembly of chip-scale packages?
A First, I would like to say that I believe for "true" CSP to be highly successful there will have to be a stronger integration between the design, fabrication, packaging and test disciplines of semiconductor manufacturing. Our existing method of manufacturing passes off the operational technology focus from one group to the next. While this may have been acceptable until now, it may not be viable for the ultimate future of CSP. I believe that the most successful efforts in CSP will come from those who have strongly aligned and integrated their design group to the fab and to packaging to provide their customers with a complete and total solution. The future of CSP depends largely on the merger of the technical disciplines we have kept at arm's length for so long.
Q Is the available equipment adequate for chip-scale packaging?
A No, I do not believe that the existing equipment base is at all adequate for CSPs. Although there are a number of "standard equipment systems" being promoted and /or specified for CSP, there is a significant amount of manufacturing process automation still needed.
To paraphrase Dr. Sassan Raissi (Expert Interview, May 1998), "technology nearly always leads the marketplace, and market demand will invariably determine the status of the equipment infrastructure." Rather then pushing forward manufacturing solutions, the equipment makers have tended to hesitate in developing equipment because they are unsure of the payback for their investment. Most suppliers want to be the second one into the market but no one really wants to be first. While this is somewhat understandable, I believe that a well thought-out design, implemented and marketed properly, can give a competent equipment maker as much as a five-year competitive advantage in the CSP market.
Q What should assemblers know about handling CSPs at the wafer level?
A Assemblers are already realizing that there is a lack of automated equipment, a need for greater process definition and more innovation required for CSP handling at the wafer level. An example is in CSP package singulation. The techniques of device singulation at wafer level have not been combined with effective unit handling or automation. This leaves the assembler to develop an automatic handling system or employ a equipment integrator who is capable of combining features from standard equipment.
Q Will there be a change in the key players from the traditional suppliers in serving CSP market needs?
A To a large degree it depends on the company itself, but I suspect the answer may be yes. I believe that as the CSP manufacturing process introduces more of the fab steps into the scenario, there will be less demand for the typical assembly equipment we know today. It may truly be easier for a fab equipment supplier to migrate backward towards a CSP solution then for an assembly equipment maker to migrate forward towards the fab process to create via systems necessary for the next generation of CSPs. Also, while the current base of assembly equipment suppliers hesitates in their development efforts, due to a lagging market demand, there will be the potential for the lesser known or second tier equipment suppliers to take advantage of the market should they so desire.
Q Is the market ready to standardize on CSP assembly equipment?
A Today it is truly difficult to standardize on CSP assembly equipment, since an overwhelming lack of CSP manufacturing standards exist. While there may be some similarities between CSP designs, footprints and process technology, we are left with a variety of interconnect methods, design outlines and materials which lend difficulty to the standardization process for assembly equipment. We can immediately standardize on our communication protocols and the software tools required for factory floor planning, of course, but the manufacturing process itself requires further definition for CSP standardization.
Considering the µBGA® process alone, the potential does indeed exist to develop a set of equipment standards which include software, hardware and process. The ability to develop standards for the equipment and automation required to manufacture the µBGA effectively (economically) can, in my mind, lead to the assembly advancement of other CSPs. If I were an equipment maker, I would carefully want to review and gauge my existing capabilities regarding the development of a fully automated µBGA assembly line. By doing so, I could very well develop significant insights into the future of the assembly equipment market and thereby gain a jump on the competition.
Q Singulation seems to be taking an increasingly important role in assembly. How can users maximize it?
A If you are referring to wafer-level CSP singulation, then of course singulation plays a vital role. We will need to reliably singulate the individual units in a process conducive towards further automation and handling. Singulation processing could take a variety of forms including, but not necessarily limited to, abrasives, water jets, lasers and combinations thereof.
However, if you are referring to the singulated processing of advanced package substrates for some CSP designs, I am not at all sure singulation is key to the future of assembly. Singulation, in this case, has come about due to poor-yielding processes. In the case of substrates, what is required is a paradigm shift in technology application towards a high-yielding process and more cost-effective solutions. This is an area of extreme importance to the industry and one which my colleagues and myself are continuing to examine.
Q Is the current level of materials used for die attach/lead bonding of CSPs suitable to these tasks?
A We are driven by performance and cost. As we continue to be performance-driven towards new packaging solutions, our processes require further assembly equipment definition and so do our materials. In our CSP developmental efforts, we will incur a variety of package integrity and reliability issues which can only be overcome through improvements in materials technology. This is, after all, the necessary evolution of technology. Pragmatically though, we continue to be driven by the eventuality of cost to performance gained. There are many technical solutions that have already been developed but are not available in the marketplace because they are not yet cost effective.