The world’s largest and most influential semiconductor companies have been cogitating about migrating to 450-mm wafer technology for years, and many of Europe's major chipmakers are still becalmed in the bay of contemplation.
Is this inert attitude encouraged by the comforting knowledge that most of today’s more important semiconductor devices are created on 300-mm wafers? If so, these companies remain in denial about a couple of plain facts regarding larger 450-mm wafers. For one, they can hold approximately twice the number of chips. Secondly, in the longer term, they provide opportunities for more cost-effective production and subsequent generation of cheaper devices.
So what’s causing such lethargy about getting 450-mm technology into production? It's mainly about the money: Researching 450-mm hasn’t come cheap, and even greater financial demands will emerge before overcoming manufacturing challenges. Perhaps that fiscal reluctance becomes more understandable.
Market Pressures
Nonetheless, relentless demand for more advanced microprocessors coupled with pressure to cut semiconductor-device costs means more companies will likely embrace 450-mm technology. That’s good news for devotees of Moore’s Law, because 450-mm production now seemingly looms as the only way to maintain that edict.
However, these revelations apparently aren’t inspiring Europe's semiconductor firms to grasp the 450-mm nettle. In fact, they’re dragging their corporate feet on this issue to the point where a real danger lurks of them being left behind.
An industry report titled “Benefits and Measures to Set Up 450mm Semiconductor Prototyping and to Keep Semiconductor Manufacturing in Europe,” compiled by industry analyst Future Horizons, states that Europe, which had led the 300-mm transition (including equipment and process R&D) via its Semiconductor 300 initiative, could have done so again at the 450-mm level.
Still, European semiconductor companies aren’t tempted to go down this path. Future Horizons’ report believes that during the 300-mm build-up in the post-dotcom-bust era, a wave of opinion grew throughout Europe that manufacturing wasn’t strategically important.
The report feels this attitude culminated in the eventual bankruptcy of Qimonda, a memory company spun out of Infineon Technologies in 2006, which at its zenith was the second largest DRAM producer globally. This collapse and subsequent closure robbed Europe of its most advanced high-volume fab. Prior to these events, STMicroelectronics’ 300-mm megafab in Catania had been shelved, with the facilities now used by 3Sun (Sharp, ENEL, STMicroelectronics) for photovoltaic solar production.
Both Infineon and STMicroelectonics have now stepped away from the memory business, highlighting a disturbing trend. High-volume semiconductor manufacturing is where Europe has particularly underperformed, preferring to move out of manufacturing to fabless models and exiting commodity mass markets in favor of niche areas. The report says these strategies have led STMicroelectronics, Infineon, and NXP to sell many of their business units.
In a somewhat depressing statement, the report maintains that in the current climate, none of these chipmakers appear willing to invest in the construction of a 450-mm fab.
Implementation of 450 mm
Looking beyond Europe, you’ll find three companies that are already marching toward the implementation of 450-mm wafer production: Intel, Samsung, and the Taiwan Semiconductor Manufacturing Company. All three are confident they will begin production using these wafers in newly equipped fabs by 2016. If so, watch out for the possible demise of 200-mm production lines.
This isn’t good news for the supposed 450-mm quiescent chipmakers of Europe or the EU in general. The region’s semiconductor industry provides the knowledge and technologies that generate about 10% of global GDP, and is perceived as an essential element of Europe's industrial community.
Riding the 450-mm bandwagon requires fiscal nerves of steel, though. Intel already confirmed that it’s begun construction of the fab D1X module 2. The new facility will be the first semiconductor manufacturing factory to process 450-mm wafers. Getting the project started has cost the company $2 billion this year, and it’s expected to take around two years.
The company also agreed to invest €829 million in Dutch lithography equipment specialist ASML’s R&D programs for extreme ultraviolet lithography (EUVL) and 450-mm wafer deployment. On top of that, it will buy €1.7Billion worth of ASML shares and invest general R&D funds totaling €3.3Billion.
Such an investment in lithography is a key strategy. EUVL is very different from the deep ultraviolet lithography used today, and presents some serious technical challenges. For instance, it must take place in a vacuum. In addition, all of the optical elements, including the photomask, must make use of defect-free multilayers, which reflect light by means of interlayer interference. These mirrors absorb around 30% of the incident light. Current pre-production EUVL contains at least two condenser multilayer mirrors, six projection multilayer mirrors, and a multilayer object (mask).
Because the optics already absorbs 96% of the available EUVL light, the ideal EUVL source will need to be sufficiently bright. Thus, EUVL source development has focused on plasmas generated by laser or discharge pulses. The mirror responsible for collecting the light is directly exposed to the plasma and, therefore, will be vulnerable to damage from the high-energy ions and other debris. Consequently, the wafer throughput of an EUVL exposure tool impinges on manufacturing capacity. EUVL requires high vacuum, so throughput is limited by the movement of wafers through the tool chamber to a few wafers per hour.
Despite all of the cost and technical hurdles faced by 450-mm technology, it’s now an industry-held perception that it will be a production reality by 2018, and will seriously affect global semiconductor sales and revenues (Figs. 1 and 2 for current trading positions).
Europe was at the cutting edge of semiconductor production during the 200- to 300-mm transition, led by the Siemens-Motorola joint venture fab in Dresden, Germany. However, the region has failed to capitalize on this technical strength, mainly due to insufficient investment. The lack of 450-mm production capability will inevitably threaten Europe’s ability to compete feasibly in the ever-advancing global semiconductor technology market.
Captions:
1. In December 2011, European semiconductor sales totaled U.S$2.78 billion. In comparison, the global semiconductor industry achieved revenue of U.S$298.3 billion worldwide in 2010.
2. European semiconductor sales increased by 0.8% in June 2013 compared to June 2012, according to World Semiconductor Trade Statistics. Q2 2013 sales rose 3.8% compared to Q1 sales. However, compared to May 2013, June European semiconductor market increased by a mere 0.1%.