Characteristics of the Semiconductor Industry
Life in the semiconductor business has been always challenging. While many of the characteristics of other industries are present, it seems like everything is measured in superlatives: the technological advances are mindboggling, the productivity improvements are astonishing, the profits are eye-popping and, of course, the product lifetimes are incredibly short, the business cycles are extreme, the losses are colossal... One wonders, however, if these extremes could continue forever as the industry approaches the respectable age of half a century. Other industries have displayed similar behaviors and time has caught up with them. The analogies would indicate that a time of consolidation and more mature, albeit less spectacular, growth and fluctuations in profit and loss as well as market should follow.
Such speculations aside, one could identify with relative ease the directions followed by the industry to maintain profitability. The driving force of this industry is the reduction of cost per function of about 30% every year. This means semiconductor devices deliver more and more performance with an increase in complexity as well as functionality of end-products.
The answer to these challenging demands of the customers was:
• Shrinking the critical dimensions
• Improving yield
• Increasing wafer size
• Improving manufacturing
The industry has been following these directions throughout its history, with the emphasis shifting from time to time. The last few years have seen, for example, a lot of focus and progress in shrinking the critical dimensions.
These shifts affect the whole industry and have always been driven by the IC manufacturers. As they develop concepts for new products with improved performance and new features required by their customers, they have to create new process technologies to build such products. In turn, the equipment and material suppliers start to develop the new manufacturing equipment and materials to be used by the IC manufacturers. This process has never been extremely efficient in the past primarily due to the lack of industry-wide standards and coordination. Traditionally the competing IC manufacturers have been protecting their advances from others and as a result it used to take years for a new technology to be adopted throughout the industry.
The problem is that developing new technologies is linked with high costs and risks. Today the cost for the development and introduction of a new technology surpasses $1.0 billion. The risk derives from the uncertainty over whether the technology will be accepted as a standard by the market. For instance wafer size has always been associated with high risks as demonstrated by the evolution beyond 100 mm: the 150 mm size was introduced earlier than expected. One reason might have been the insufficient productivity gain on 125 mm wafers compared to 100 mm wafers, represented by the area increase (Figure 1).
Figure 1. Wafer size history.
This illustrates the high risk of driving a wafer size conversion if all the factors like chip volume demand, cost advantage and general market conditions are not aligned. It is actually most important for the equipment and materials suppliers because they need to establish economies of scale by choosing the right technology and standards. On the other hand the competitive advantage for the companies that have made the right decision in the standardization race for a given technology is generally short-lived (Figure 2).
Figure 2. Lifetime of wafer size generations.
This complex situation leads to the important question that we want to discuss in this paper: how to develop a new production standard for the next generation manufacturing technology in the most effective way?
Effective Solutions for Future Technology Challenges
Semiconductor manufacturers differentiate each other by product features and performance, by design and product verification cycles and by manufacturing capabilities. They all strive to translate such differences in cost advantages and therefore increased profits. Introducing standards in such an environment is not easy. However the high and galloping increase in the cost for technology and manufacturing has recently forced the manufacturers to look at more effective approaches.
One approach has been the formation of consortia focused on the development of standards for new equipment and manufacturing technology. An example is I300I, the International 300 mm Initiative, which is now part of International SEMATECH.
Yet another possibility has been the formation of strategic alliances with the purpose of sharing knowhow and reduce the costs and risks. Some of them have been focused on new product and processing technology. Successful examples include the Motorola–IBM alliance for the development of a new microprocessor architecture and the Siemens–Toshiba-IBM alliance for the development of advanced DRAM technology.
One can look at alliances in two different dimensions: one is concerned with what the partners can achieve together, the second with what the partners can gain for themselves. From this standpoint alliances create value by improving the competitive standing of the partners against others, and on the other hand the partners attempt to extract value for themselves from within the alliance. This dual process of value creation and value extraction in alliances can be thought of in terms of baking and then sharing a pie with each partner contributing different resources and skills. Cooperation of this kind in principle offers a lot of advantages. But as some empirical studies show strategic alliances are not successful per se. The studies report a failure rate between 60% and 70%.
Designing a Successful R&D Alliance
The case for the formation of a strategic alliance to develop the next semiconductor manufacturing technology is pretty straightforward. Firstly, such an endeavor, when undertaken in the past by one company has resulted in enormous costs for the company and overall delay in the industry adoption and readiness for the technology. Secondly, the human capital, costs and risks involved would seriously stretch the resources even of one of the major semiconductor houses. Lastly, the development of such a technology by an alliance rather than a single partner would increase the credibility of the finished product and facilitate the adoption of standards.
Once Motorola and Siemens have independently decided to pursue the development of the 300 mm manufacturing technology in an alliance, several considerations had to be addressed:
• Was it strategic enough for both the companies?
• Was there a mutual benefit?
• Did synergies exist between the two companies?
• Was there a strategic fit (compatible and/or complementary) between the two companies?
Inasmuch as the goal was to develop and introduce the next semiconductor manufacturing technology, such an endeavor was clearly strategic for both companies. The mutual benefits were also obvious: Motorola already had a 300 mm development program under way, while Siemens had established a leading edge manufacturing line at Dresden, where a reference advanced 200 mm manufacturing operation was in place.
The synergies and strategic fit have already been demonstrated in another successful alliance of the two companies, White Oak Semiconductor, a 200 mm manufacturing line in Richmond, Virginia. In addition the project has benefited from funding from the German Federal Ministry of Science and Technology (BMBF) as well as from the State of Saxony where the city of Dresden is located (Figure 3).
Figure 3. Factors driving SEMICONDUCTOR300 as a Joint Venture.
One of the major issues in a successful cooperation is to establish trust between the two partners. In the case of Motorola and Siemens this was not really a problem because of the joint manufacturing activities since 1997 at White Oak Semiconductor. Trust between the two companies existed before the partnership discussions started.
Developing a new production standard within an alliance raises first of all a managerial challenge. Such an approach requires not only a technological but also a considerable managerial effort. As the literature and experience indicates most failures of alliances are caused by managerial problems. Many companies mismanage their alliances, leading to wasted money, time and effort. The dimensions of the managerial problems were apparent from the beginning in the light of different geographic locations, different firm cultures, different ethnological backgrounds.
The Managerial Challenge
In February 1998 Motorola and Siemens established SEMICONDUCTOR300 as a 300 mm technology development and manufacturing pilot line. The location was Dresden, Germany where Siemens had already a large volume 200 mm manufacturing facility. The choice of the location was intentional as it was capable of providing a direct and immediate comparison of the products from the 300 mm pilot line with those from the established 200 mm line. A number of managerial challenges were laying ahead:
• Define the legal framework of the enterprise.
• Identify employees from both companies that showed willingness and ability to work in an ambiguous, unfamiliar, cross-functional, trans-cultural environment.
• Establish a clear and compelling vision/mission for the enterprise.
• Establish an appropriate organizational and communication structure.
• Define and communicate clearly the project goals.
• Establish standards for the project reporting and documentation.
• Define what constitutes success, how to measure it and the appropriate incentive plans.
• Impact of the Legal Framework. Very important besides the legal agreement between the parent companies and the disclosure rights to the development results was the legal form of the company. The country specific legislation dictated whether a joint project can be established in the form of a cooperation or a new company has to be established, as in the case of SEMICONDUCTOR300. The requirement to establish a new company had implications not only from the financial and legal point of view, but also for the employees. From their parent companies prospective the employees represent “the parent company” in the JV; on the other hand legally and in order to establish a new company culture, they identified themselves with the new legal entity.
• Human Resources. The technical challenge undertaken by the enterprise dictated the need for very experienced and highly innovative employees. Given the local conditions, many of these employees have been recruited from other parts of Germany, other countries in Europe and the United States. Dresden is located in the so called “new states” in the former eastern part of Germany. This added an additional dimension because of different living conditions and cultural background of this part of the country. It is true that after the peaceful revolution the people in eastern Germany have experienced an unprecedented change in their economic status and personal values. Nevertheless, for the employees coming from outside Dresden an appropriate working environment and adequate living conditions for their families had to be established. Specific issues related to the diverse background of the work force had to be addressed through the use of specialized service providers. Over 450 people are expected to be employed by the enterprise by the end of next year.
• Clear Vision/Mission. The mission of SEMICONDUCTOR300 is to achieve a breakthrough in semiconductor manufacturing by developing the next generation manufacturing technology capable of delivering a cost reduction of 30%-40% per chip compared to 200 mm. The detailed deliverables for this goal had been defined, the main assets and future benefits identified and detailed timelines for the execution of the mission have been set. The vision/mission as well as the deliverables, timelines and progress thereof have been widely communicated to the employees using audio-visual means as well as regular “all hands” type of meetings.
• Organizational and communication structure. A structure that blended characteristics of typical structures in the parent companies has been designed. It was important for the structure to satisfy the needs in both the initial heavy developmental phase and the subsequent pilot line phase. In order to ensure that the interests of the parent companies would be adequately represented and a balance would be maintained, an office of the General Manager with one representative from each company was installed as the leadership of the enterprise.
The notion that people need to communicate more is perhaps the most widely accepted idea in management, indeed in all human relationships. In fact most organizations are overcommunicating: meetings, conferences, memos, phone calls and electronic mail overwhelm managers and employees. Increasingly, we seem to believe that everybody should be in everything. In the special circumstances of a multicultural cooperation, the form of communication is sometimes more important than content – the power of the invisible. Meeting and communication guidelines have to be set, understood and accepted by all partners in the same way in order to become effective.
• Project goals. The project has been funded with approx. $1.0 billion for its three year duration. Given the complexity of the task, the large investment and the size of the enterprise, a strict annual goals-setting process has been implemented. The goals have been selected to support the proper definition and execution of a business plan and in the same time to ensure congruence with the company mission and the achievement of the deliverables in the timelines committed. The communication of clear goals and the quarterly review of the progress towards their achievement have kept the whole organization focused and the project execution on track.
• Reporting and documentation. The JV Agreement has specifically defined documentation requirements. These have been specified in great detailed and provided an excellent framework for capturing the data generated during the project. A number of metrics to measure the progress of the documentation have been set and reviewed on a regular basis. The form and content has to be reviewed by the parent companies in order to guarantee adequate fulfillment of the parent companies expectations. These funding entities have clear progress reporting requirements and by fulfilling them the enterprise has implemented high levels of project results documentation.
• Definition and measurement of success. Utmost attention has been paid to setting project deliverables as well as annual goals that were specific, measurable, achievable, relevant and time-bound. As mentioned before, a regular review process allowed early detection of issues and potential delays and timely implementation of corrective actions. Since the project overall goals and timelines were extremely aggressive it was important to establish an appropriate incentive system for all the employees. This consisted of a combination of team and individual goals, a structured measurement system and biannual reviews. Individual bonuses could become significant when both the team and individual goals were achieved.
Results to Date and the Challenges of the Next Phase
The project experienced a good start with the goals defined, the organization in place and the essential players present or committed within the first six months from inception. The first hurdle – the acquisition of 300 mm equipment – was passed successfully with a complete process tool-set in place by November 1998. This in itself was an important achievement given the fact that it was achieved right in the middle of a “double dip” recession that has hurt all the industry players very seriously, especially the equipment suppliers.
However, the crucial test was the demonstration of the capability of this equipment set. This was also achieved when in December 1998 the very first lot of 300 mm wafers processed entirely at SEMICONDUCTOR300 yielded fully functional 64M DRAM chips. It was an outstanding success achieved in less than one year from the formation of the enterprise!
Key to this achievement were the daily heroics and high motivation of the entire team, the unwavering support of the parent companies and the excellent cooperation of all the equipment and material suppliers. As a side benefit from this extremely intense period a spirit of strong cooperation and trust between the team members emerged.
The following phase was the ramping up of the pilot line and the qualification of the product. The focus shifted from equipment to process in order to achieve the necessary yields and reliability results for a successful qualification. To facilitate this the partners chose not to start with the most advanced process technology on a new wafer size, as it has been the approach in the past. Rather, an advanced technology already in volume production – 0.25 µm – and a successful product with a good, solid history – 64M DRAM – have been chosen. This allowed for an easy comparison of the same product manufactured on 200 mm and 300 mm wafers. This was one of the key elements of the concept as it allowed the identification of the equipment and wafer size related issues before shrinking down the process technology to the feature size required for volume production. The manufacturing issues became apparent as it was to be expected in a pilot line with mostly one-of-a-kind equipment and no manufacturing redundances built in. In spite of this the ramp was an overall success and the yields achieved surpassed expectations by 200%. The final success was the passing of the stringent qualification requirements in September 1999 and therefore the ability to ship products to the customers.
Wide recognition of these successes further increased the team confidence and cemented the trust and spirit of cooperation. It also constituted, together with the accomplishment of several essential milestones, a measure of the success of the alliance idea.
The next phase will see the main focus return to the equipment. The alliance’s partnerships with the various equipment suppliers will concentrate on equipment optimization in order to prepare second generation tools with the right manufacturing and technology performance at the right cost for a high volume 300 mm wafer fab. The alliance will assess the tools readiness status as a precondition for the parents decisions concerning the building of volume manufacturing fabs. The managerial challenge during this phase will be to keep the organization from unravelling due to excessive segmentation of these tasks or due to sometimes conflicting goals of the parents.
Lessons Learned and a Look into the Future
The successes obtained so far by SEMICONDUCTOR300 indicate that strategic alliances – Joint Ventures in particular – are effective ways to address the increasing costs and risks associated with the development of new technologies in the semiconductor industry. Indeed one may estimate that a potential cost saving of up to 5 times was achieved over the previous new manufacturing technology effort.
Identifying and establishing the right solutions for the managerial challenges from the beginning has enabled the team to stay focused on the tasks and work effectively to meet the goals. A few important lessons should probably be mentioned:
• Establish early a clear and compelling vision, capable of energizing the team and conveying a powerful message to all the alliance’s constituencies: parent companies, suppliers, political/governmental institutions, international institutions, other semiconductor manufacturers and media.
• Establish clear measurements of success accepted in the different company cultures and a disciplined goals review process. Give team members a stake in the success of the enterprise by instituting a meaningful and rational incentive program that is not just copying the concept of one or the other partner.
• Allow for the cultural differences to coexist rather than trying to create a completely new culture, especially if the alliance is only established for a limited time. Try, however, to build as many bridges as possible over cultural barriers.
• Show flexibility in priorities, methods, organization, etc as long as it serves the purpose of the enterprise. If the alliance is set up only for a relatively short duration, the focus has to be on development results and achieving the goals in time and within budget. Naturally, such a situation only increases the management challenge. One has to establish a creative and synergetic working atmosphere to compensate for the lack of long term perspective!
• It is the belief of the authors that alliances will become more and more a way of the future, perhaps the only way, for addressing the increasingly stringent requirements of the technology development in the semiconductor industry. The learning from successful joint ventures like SEMICONDUCTOR300 could help future alliances to optimize their chances of success and hopefully improve the percentage of successful alliances overall.
 Yves L.Doz, Garry Hamel: Alliance Advantage – the art of creating value through Partnering Harvard Business School Press, Boston MA 1998
 Larraine Segil : Intelligent business alliances – how to profit using today´s most important strategic tool; Times Book by Random House Inc., New York 1996
Horia A. Grecu is the General Manager of SEMICONDUCTOR300, together with Dr Peter Kuecher. Prior to this he was General Manager of MOS10, a Motorola IC manufacturing plant in Irvine, California. Over the last 20 years Mr Grecu held various positions, both in engineering and production, with Motorola and other IC manufacturers. He has a MBA from the University of California, Irvine and a MS in Physics from the University of Bucharest, Romania.
peter Kücher, PhD is the General Manager of SEMICONDUCTOR300, together with Horia A. Grecu. Prior to this he was the Director of the 300 mm program for Siemens Semiconductors (now Infineon Technologies). Dr Kuecher held various positions in process engineering with Siemens and managed several European projects within ESPRIT and JESSI. He was also the Siemens project manager in the TRIAD project with IBM and Toshiba. He has a PhD in Applied Physics from the University of Regensburg, Germany.