With the Starlith 1700i immersion objective from Carl Zeiss SMT the chip manufacture is ready for 45 nm series production. (Source: Carl Zeiss SMT, approved)

The miniaturization of integrated circuits will continue over the next 15 years. It is already possible to mass-produce structures measuring only 45 nm. Now the plan is to produce structural widths of 32 nm. This will make it possible to produce microprocessors that are 100-1000 times more powerful than the ones at present. However, this will require a fundamental technological shift.

Since the early 1970's chip manufacture in the semiconductor industry has followed the so-called Moore's Law: the clock speed of microprocessors doubles every two years. The International Technology Roadmap for Semiconductors (ITRS) and leading semiconductor manufacturers estimate that the miniaturization of integrated circuits will continue over the next 15 years. Optical lithography has a leverage effect of 1:100 - 1000 on the further processing industry.

For this reason, companies as well as research centers are increasing their efforts to develop technologies for the mass technical production of the smallest possible mask structures. On February 27, 2008 the leading company in this field, Carl Zeiss SMT in Oberkochen and its partner company, the manufacturer of semiconductor exposure devices ASML in Veldhoven, Netherlands, presented the latest state of development in prestigious lectures at the traditional optical colloquium at the University of Stuttgart.

Currently, semiconductor manufacturers are using ultraviolet light with a wavelength of 193 nm (a nanometer is a millionth of a millimeter), in order to project structures measuring 45 nm via lenses onto a silicon wafer coated with photoresist. The sensitive photoresist coating is then developed, the non-exposed areas are etched away and the new semiconductor materials are applied. The process is repeated up to 40 times with different masks with a degree of precision so that the order of magnitude is below the exposed structural width. This technology was made possible through the use of "Starlight 1700i" immersion objectives for the series production, which were first developed by Carl Zeiss SMT at the end of 2007 as a technical challenge. A water film between the objective and the silicon wafer to be structured makes it possible to expose wafers with a diameter of only 300 mm. "Starlith 1700i" offers a 30% higher resolution and enables the series production of microchips with 45 nm structures - a first. The objectives are only 1500 mm thick and consist of a lens system made of special glass with a diameter of over 300 mm, which is still permeable for a wavelength of 193 nm (Figure: Starlith 1700i light objective). The requirements for the optical and mechanical components in terms of precision required are very strict: they have to be 1-2 orders of magnitude below the generated structural width of 45 nm. This requirement necessitates ultra-precise measurement and a resulting simulation of the overall system performance. For this development Carl Zeiss SMT won the German Industry award for innovation in the major company category in 2007.

For economic reasons the semiconductor industry requires even smaller structures of 32 nm and below. That requires a completely new exposure process: Extreme Ultra-Violet Lithography (EUV). This EUV technology which cannot yet be fully implemented in technical terms operates with a 13 nm wavelength. Reducing the wavelength by a single order of magnitude however requires a level of precision of the manufacturing technology for the objectives and mask exposure which is another order of magnitude below the one currently being used. This is necessitating a fundamental technological shift as glass is opaque at this wavelength. High-precision mirror optics has to take over the formation of the exposed wave front. The variations of the aspherical mirrored surfaces and the surface roughness must be under 0.1 nm, which is already in the atomic range. The complete exposure system, consisting of several precisely coordinated mirrors is operated in a high vacuum because even the smallest quantities of air can absorb the EUV radiation.

The transition to 13 nm wavelength, however, will also require new resists, new exposure systems for the mirror optics and new sources of radiation. Here companies and research institutes such as the Fraunhofer-Institut für Angewandte Optik und Feinmechanik (IOF) [Fraunhofer Institute for applied optics and precision engineering] in Jena are working on the development of special coating processes and measuring processes to characterize the new components and the development of high-precision joint and assembly processes. The prototype of a laboratory resist exposure device is being constructed for the development of photoresists for EUV lithography. For this the IOF team won the Thüringer Research Award 2007 in the category of "Applied Research" on January 9, 2008. For the exposure a high-performance plasma source is conceivable which must have a radiation performance of about 1 k watt and this does not yet exist. The lifetimes of source and exposure optics currently represent the biggest challenges for the transition of EUV lithography from development to production.

The end of optical lithography has often been prophesized for 2010. However, complex technical possibilities are constantly being found to push back this limit and it is currently 2016.  Experts agree that it is important way to go down the route of smaller structures because ultimately this will make it possible to manufacture microprocessors that are 100 - 1000 times more powerful than the ones today. In particular the manufacturers of processors are interested in higher pack densities, more integrated functions and in reducing the power consumption for consumer products. The semiconductor industry is putting pressure on manufacturers: in the near future Intel plans to present a processor measuring 45 nm with a power consumption of only 50 watts and from 2009 plans to start production of 32 nm processors. These new circuits will drive forward the overall progress of technology and will benefit a wide range of everyday applications.