EUV Lithography

EUV Lithography Proof Is in the Printing

An old adage says, "If you can't measure it, you can't make it." So it's no accident that metrology beamlines built and operated at the ALS under the auspices of Berkeley Lab's Center for X-Ray Optics (CXRO) have been instrumental in a 5-year, $250-million industry–national laboratory effort to bring extreme ultraviolet (EUV) lithography to the commercial stage. EUV lithography is the future chip-printing technology that the Semiconductor Industry Association began backing in 2001 as the likely successor, around the year 2007, to the reigning family of refractive optical lithography techniques. The EUV promise is that with wavelengths 50 times smaller than those of visible light, it will be able to draw circuit patterns just tens of nanometers wide. In comparison, the current industry state-of-the-art chips have patterns with 130-nm lines.

Thanks in part to the CXRO program, interferometry is now routinely used for the characterization and alignment of optics for EUV lithography, but the ultimate performance metric remains the quality of the printed patterns. Moreover, the comparison of a lithographic image with that predicted from interferometry-derived wavefront maps is also useful for verifying and improving the predictive power of wavefront metrology. To address these issues, CXRO added small-field printing capabilities to the EUV phase-shifting/point diffraction interferometer (PS/PDI) on ALS Beamline 12.0.1. The first printing results reveal the extraordinarily high quality of the lithography optic and in doing so demonstrate the success of CXRO's EUV metrology beamlines in helping to achieve very tight specifications for figure, finish, and multilayer coatings of the aspherical mirrors comprising the lithography optics.

The scanning mirror directs focused EUV light from the Kirkpatrick-Baez mirrors into the converted interferometer with the desired degree of spatial coherence and illumination pattern. Installed inside the interferometer tank, the Set-2 optic images the EUV light reflected from the mask-carrying reticle onto a resist-covered silicon wafer.

How to Obey Moore's Law

In conformance with Intel co-founder Gordon Moore's 1965 prediction, now known as "Moore's Law," the density of circuit elements on microchips has doubled roughly every 12 to 18 months for more than 30 years, resulting in ever smaller, faster, and cheaper computers. However, manufacturers know that the traditional technique for printing circuit patterns—optical lithography based on refractive optics (lenses)—cannot continue indefinitely on this course. Today's leading candidate for a successor, known as EUV lithography, relies on reflective optics (mirrors) to image patterns from masks onto the surface of a silicon wafer that will ultimately be diced into microchips. The first computer processors produced with EUV technology beginning around 2007 are expected to be almost ten times faster than today's most powerful chips, and the storage capacity of memory chips will increase even more. But before that day arrives, there is the matter of producing accurate EUV lithography cameras. Naulleau et al. have verified that the optics fabricated for a prototype camera are indeed on the path to the required performance by using them to print actual test patterns with ultrathin line widths.

In the printing configuration, the test station is referred to as the static exposure station (SES). Static means that the scanning and stepping systems required to print complete circuit patterns on all the chips on a silicon wafer are not implemented. Instead, only small ("microfield") test patterns are imaged. Key features of the SES are the ability to control the coherence of the illuminating EUV light (partially coherent is optimum for printing) and the illumination pattern (disk-shaped, annular, dipolar pairs of disks, channel-shaped, etc.). The combined PS/PDI–SES system remains extremely flexible in that switching between the interferometry and printing modes can be accomplished in approximately two weeks.

Two 4x reducing optical systems have been developed as part of the EUV lithography program in which Intel, Motorola, Advanced Micro Devices, IBM, Micron Technology, and Infineon Technologies (the EUV Limited Liability Corporation) partnered with the "Virtual National Laboratory," consisting of groups from Lawrence Livermore National Laboratory, Sandia National Laboratories, and Berkeley Lab. The second, much higher quality, Set-2 optic is destined for integration into the prototype printing machine (the Engineering Test Stand or ETS) for full-field scanned imaging, but the CXRO team has already obtained valuable information by using the SES to lithographically characterize the static imaging performance of the Set-2 optic.

Printing up to specs. Designed for printing patterns with sub-100-nm features, the Set-2 optic imaged elbow test patterns with a line-to-spacing ratio of 1:1 with high fidelity down to line widths of 70 nm.

Designed to image features of 100 nm and below in patterns with a dense 1:1 ratio of line widths and line spacing, the ETS Set-2 optic easily lived up to its specs, achieving line widths as narrow as 70 nm in elbow patterns. By adjusting the illumination pattern and the exposure dose, the team printed less densely spaced lines with widths down to 39 nm. These results indicate that with the new optic set expected to arrive at the ALS for testing in November 2002, it should be possible to print features in the 16-nm to 18-nm range and thus meet the production requirements set for chips with 1 billion transistors and up in the years 2007 to 2010.

By controlling the exposure, the CXRO team was able to print line widths down to 39 nm with the Set-2 optic for an elbow test pattern with a line-to-spacing ratio of 3:1

Research conducted by P.P. Naulleau, K.A. Goldberg, E.H. Anderson, D. Attwood, P. Batson, P. Denham, E. Gullikson, B. Harteneck, B. Hoef, K. Jackson, D. Olynick, S. Rekawa, and F. Salmassi (Berkeley Lab); J. Bokor (University of California, Berkeley, and Berkeley Lab); K. Blaedel, H. Chapman, L. Hale, R. Soufli, E. Spiller, D. Sweeney, J. Taylor, and C. Walton (Lawrence Livermore National Laboratory); G. Cardinale, A. Ray-Chaudhuri, A. Fisher, G. Kubiak, D. O'Connell, R. Stulen, and D. Tichenor (Sandia National Laboratories); and C.W. Gwyn, P.-Y. Yan, and G. Zhang (Intel Corporation).

Research funding: Extreme Ultraviolet Limited Liability Corporation and U.S. Department of Energy, Office of Basic Energy Sciences (BES). Operation of the ALS is supported by BES.

Publication about this research: P.P. Naulleau et al., "Static Microfield Printing at the Advanced Light Source with the ETS Set-2 Optic," Proc. SPIE 4688-05 (2002, in press).

ALSNews Vol. 202, July 3, 2002

More ALS Science