Europeans collaborate toward next-generation lithography
Extreme-UV lithography has become the most promising next-generation lithography technology for reaching the 50-nm node and beyond in integrated-circuit feature sizes, according to International Sematech (Austin, TX).
Extreme-UV lithography has become the most promising next-generation lithography technology for reaching the 50-nm node and beyond in integrated-circuit feature sizes, according to International Sematech (Austin, TX). Intense light at 13.5 nm for extreme-UV lithography (EUVL) is no longer produced by lasers but by either discharge-produced plasma (DPP) or laser-produced plasma (LPP) sources. The unprecedented expense and difficulty of realizing such sources constitutes the most critical issue for the development of EUVL production tools. Consequently, the enormous challenge for photostepper manufacturers, such as ASML (Veldhoven, Netherlands), Nikon, and Canon (both Tokyo, Japan), is to avoid the so-called "Concorde" paradox, in which amazing, high-end technology—as in the case of the famous airplane—simply becomes too expensive to be economically viable.
To avoid reliving the Concorde story in the semiconductor industry, EUVL production rates must meet or exceed one hundred 300-mm wafers per hour. Achieving that based on current estimates requires 115 W of useful EUV power to run a production tool. Taking into account all of the EUV losses due to propagation and filters, this calls for up to 500 W of EUV light at 13.5-nm in 2π steradian within a 2% bandwidth. Despite impressive worldwide R&D efforts on EUV lithography for almost a decade, current technology simply cannot provide such tremendous source powers, according to statements made last October at the 2nd International EUVL Symposium held in Antwerp (Belgium). For both LPP and DPP sources, the best performances are still about an order of magnitude below the required powers.
The main laser-produced plasma sources in Europe are from left to right, the Exulite source, the Xtreme source, the Powerlase source, and the IOE source.
Although most source suppliers prefer DPP technology because of its relative simplicity and more efficient conversion of electrical into EUV power, both concepts remain in the race. The success of LPP technology will be strongly linked to the emergence of an affordable and efficient laser technology to excite the emitting plasma (see figure). Following the Antwerp symposium, introduction of EUVL has been delayed another two years to 2009.
In recognition of the importance of EUV lithography, several European countries have established research and development networks. Most EUVL-related research is done within the Microelectronics Development for European Applications (MEDEA+), which describes itself as the "program for advanced cooperative R&D in microelectronics to ensure Europe's technological and industrial competitiveness in this sector on a worldwide basis."
Extreme-UV lithography is only a part of MEDEA+ activities, which include 35 projects covering the entire domain of microelectronics, with 2600 scientist and engineers from 220 MEDEA+ partner companies and institutions in 17 European countries. The budget allocated for the eight-year duration of MEDEA+ is _4 billion (about US$4.86 billion), and 25% of it is devoted to EUVL.
Currently, four three-year EUVL-related MEDEA+ projects involving 30 partners plus subcontractors from seven countries are under way, for a total amount approaching 2000 man-years. Since the beginning, all projects have been strongly linked to each other in a collaborative effort to "win the battle for the next-generation lithography solution." The EUVL activities are coordinated by the Cluster Steering Council presided over by ASML. In addition, within the 6th Framework Program an Integrated Project, called "More Moore," unites the main European research efforts devoted to EUV lithography. And the overall effort is bearing fruit.
At the time of this writing Exitech (Oxford, England) was in the process of delivering an EUV microstepper to International Sematech before the end of 2003, equipped with a DPP source built by Xtreme Technologies, a joint-venture between Lambda Physik (Göttingen, Germany) and Jenoptik (Jena, Germany). Xtreme is actually the only company in Europe developing both source concepts in parallel, that is, a Z-pinch type DPP source as well as an LPP source based on liquid µ-jet target and a master-oscillator power-amplifier (MOPA) laser scheme. Both sources use a xenon emitter, but tin-based targets are also being considered. According to Extreme, DPP sources may be a viable solution for EUVL in the early stages, whereas LPP sources are better suited for future EUVL high-volume tools.
Another major European source developer within MEDEA+ is the Exulite consortium in France that coordinates joint efforts of the national laboratory CEA (Saclay, France), Alcatel Vacuum Technology (AVT; Annecy, France) and Thales Laser (Orsay, France). Under the leadership of AVT, a modular LPP source with 50-W output power is being developed for mid-2005. The adopted source concept involves a strongly confined cryogenic xenon jet target and the use of 10 identical power-oscillator modules that are based on continuous wave (CW) diode-pumped solid-state technology. Initially, this very reliable, efficient and compact laser technology had been developed for laser-induced isotope separation (SILVA project). The modular LPP concept for EUVL sources was first proposed by CEA and Thales in 2000 at the 2nd Sematech EUVL-source workshop held in San Francisco, CA. The highly redundant Exulite source approach enables dramatic cost reductions for development as well as future production.
Powerlase (Crawley, England), a source developer outside the MEDEA+ circle, has followed a philosophy similar to Exulite in developing a modular LPP source. During the symposium in Antwerp, Powerlase statements indicated that the LPP source price is expected to drop dramatically to become comparable to that of DPP sources. This is expected mainly because of rapidly decreasing prices of high-power laser diodes along with rapidly increasing production volumes.
European source development activity concerning DPP sources is equally rich within the MEDEA+ project. For instance, Philips EUV (Aachen, Germany) is developing a high power source based on hollow cathode geometry. The Philips program includes basic theoretical and experimental research, engineering related problems, metrology and source-to-stepper integration studies. Currently, the realization of a set of alpha-tool sources is under way. Aixuv (Aachen, Germany) is already commercializing DPP sources for metrology.
Also in the domain of sources for metrology, the Institute of Optoelectronics (Warsaw, Poland), within the Military University of Technology is developing an LPP source devoted to EUVL inspection applications. And the "Flying Circus" of the FOM Institute (Utrecht, The Netherlands), has become a reference among all the EUV source calibration tools.
European projects such as MEDEA+ have recently become the main pathways for arriving at the ambitious goal of developing industrial tools for EUVL both for production and inspection. Only a small number of EUV-source providers are expected to make it over the high technical and economic hurdles that loom in the near future but there is certainly a reasonable probability that some of those who do make it will "speak European."
TIBERIO CECCOTTI is a research scientist currently assigned to the Exulite Project at the Groupe d'Application des Plasmas, Commissariat à l'Energie Atomique, CEA/Saclay, SPAM, Gif s/Yvette, France; e-mail: firstname.lastname@example.org.