Ultrahighaccuracy optical testing creating diffractionlimited shortwavelength optical systems

Ultra-high-accuracy optical testing: creating diffraction-limited shortwavelength optical systems Kenneth A. Goldberg Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

Ultra-high-accuracy optical testing: creating diffraction-limited shortwavelength optical systems Kenneth A. Goldberg Patrick Naulleau, Senajith Rekawa, Paul Denham, J. Alexander Liddle, Keith Jackson, Erik Anderson, K. Bradley, R. Delano, B. Gunion, B. Harteneck, B. Hoef, G. Jones, C. D. Kemp, D. Olynick, R. Oort, F. Salmassi, R. Tackaberry Center for X-Ray Optics, Lawrence Berkeley National Laboratory in collaboration with J. Taylor, G. Sommargren, H. Chapman, D. Phillion, M. Johnson, A. Barty, R. Soufli, S. Bajt, et al. Lawrence Livermore National Laboratory and the EUV LLC and VNL KAGoldberg@lbl. gov, Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16 SPIE 2005, 5900 -16 K. Dean, et al. International Sematech

Consider two state-of-the-art telescopes SDO — Solar Dynamics Observatory AIA Normal-incidence mirrors Angular Resolution: 0. 6” per pixel Please see Soufli, et al. SPIE 5901 -24 (Mon) SDO / AIA material courtesy of LMSAL CHANDRA X-Ray Observatory Nested glancing-incidence mirrors Angular Resolution: ~0. 5” in 0. 5– 10 ke. V Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16 NASA / CXC / SAO

Better optics / Better angular resolution is possible Resolution is set by design compromises • Telescope size • Detector-pixel size • Fabrication limits • Collection efficiency • etc. However, thanks to the semiconductor industry, technology for much higher quality lenses is now available in the EUV. So, what does the semiconductor industry want with EUV lenses? Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

Lithography follows Moore’s Law DRAM half-pitch Intel plans industry plans 45 nm 2007 – 2010 32 nm 2009 – 22 nm 2011 15 nm 8. 6 GB 2013 – 2013 2016 – DRAM 34 GB 69 GB EUV 2019 By 2009– 13, mass production of lithographic-quality EUV lenses. Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

Diffraction-limited EUV optics for photolithography EUV projection lenses may be the highest quality optical systems ever produced = 13 nm, ~90 e. V Mo/Si multilayer-coated for near-normal incidence: R ≈ 70% (Lawrence Livermore, Lawrence Berkeley, et al. ) Diffraction-limited spatial resolution MET projection lens Up to 0. 3 NA, ƒ/1. 67 Rayleigh resolution: 1. 22 / NA 27 -nm half-pitch Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16 Courtesy J. Taylor, LLNL

Industry-funded EUV projection lithography research at Berkeley, Livermore, Sandia Nat’l Labs patterned reflective mask resist-coated wafer Engineering Test Stand Sandia National Labs. Courtesy, Bill Replogle Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

Reaching “diffraction-limited” performance Rayleigh criterion /4 P-V 3. 35 nm Maréchal criterion /14 rms 0. 96 nm ~ /50 rms 0. 27 nm Lithographic criterion If you can measure it, you can make it. Ultra-high-accuracy optical testing is the key Visible-light and EUV interferometry with sub-Å RMS accuracy. Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

Inaccurate interferometry cost NASA $Billions Before After Hubble Space Telescope Kenneth. Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16 Kenneth Hubble repair, 1993 COSTAR optic installed

Comparison of slope and roughness requirements EUV Lithographic Optics SDO — AIA instrument Goal Slope error 0. 5 arcsec ≤ 5 µrad full aperture Roughness Goal 32 -nm half-pitch Slope error 37 -Zernikes ≤ 4. 4 Å MSFR 1/ƒ = (4 µm, 4 mm) mid spatial-freq. Micro-roughness ≤ 4. 4 Å HSFR 1/ƒ = (9 nm, 4 µm) high spatial-freq. 0. 3– 1. 0 µrad measured ≤ 1– 2 Å The specs are several times tighter, and achievable Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

Numerous factors contribute to the wavefront Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

Pushing visible-light interferometry Livermore scientists developed the PSDI or Sommargren Interferometer Single Mirror • Two beams are launched into a fiber with a time delay. • A pinhole in a mirror creates the test and reference beams. Complete System • One fiber + pinhole at each conjugate. G. Sommargren, J. Taylor, M. Johnson, D. Phillion, H. Chapman, A. Barty, et al. (LLNL) (SPIE 5869– 28, 30) Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

High sensitivity to multilayer properties Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

EUV interferometers used at the ALS Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

EUV interferometers used at the ALS — PS/PDI Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

EUV interferometers used at the ALS — LSI Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

the illuminated MET pupil EUV Light transmitted light Kenneth. Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

PS/PDI interferogram EUV Light ultra-high accuracy Kenneth. Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

shearing interferogram EUV Light efficient measurement method Kenneth. Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

A long track record of EUV Interferometry, alignment optimization and imaging at LBNL (since ’ 93) Berkeley 10 x. I 10 x. A 10 x. B 2 F 2 X ETS Set-1 ETS Set-2 MET 2 -mirror, 10 x Schwarzschild objectives 4 -mirror, 4 x ETS projection optics 2 -mirror, 5 x MET optic NA ≥ 0. 08 ƒ / 6. 3 NA = 0. 1 ƒ / 5. 0 NA = 0. 3 ƒ / 1. 67 higher resolution time Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16 higher quality

ETS Projection Optic: off-axis, large field Work sponsored by the EUV LLC ~1. 1 -m mask-to-wafer M 2 M 4 M 3 M 1 Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

The 0. 3 -NA Micro-Exposure Tool: high resolution Work sponsored by International SEMATECH Courtesy of J. Taylor, LLNL Illumination Mirrors: Zeiss MET NA = 0. 3, ƒ/ 1. 67 = 13. 4 nm 5 x demag. 200 x 600 µm field capable of ≥ 12 -nm printing Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16 Coating & Assembly: LLNL

MET at-wavelength interferometry and alignment MET Micro-Exposure Tool Wavefront measurement during alignment central field point astig coma sph ab trifoil h-o s. RMS 0. 04 nm 0. 06 nm 0. 04 nm 0. 14 nm 0. 37 nm 0. 55 nm /24. 5 • EUV interferometry maps the 3 -D field of view. • Alignment sets astigmatism, coma, spherical aberration arbitrarily small. Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

Small tolerances necessitate ultra-high accuracy one nm /24. 5 = 0. 55 nm wavefront quality of best EUV optic to date /50 = 0. 27 nm EUVL -tool, required wavefront quality /135 = ~100 pm EUV interferometer accuracy @ 0. 3 NA /330 = 40 pm EUV interferometer accuracy @ 0. 1 NA /255 = 53 pm Bohr radius, a 0 Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

The keys to achieving ultra-high accuracy 1) Innovative calibration methods (null tests) 2) New interferogram analysis techniques for minimizing phase-measurement errors 3) High-quality “reference” pinholes (Nanowriter) We isolate and measure geometric & systematic error sources so they can be subtracted. Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

two-pinhole null-test interferogram system calibration for high accuracy Kenneth. Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

grating null-test interferogram system calibration for high accuracy Kenneth. Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

Developing state-of-the art pinholes for spherical reference-wave accuracy TEMPEST-3 D Modeling Nanofabrication (Nanowriter) Pinhole-array diffraction TEM image pinhole Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16 Intensity 100 nm diffraction angle Intensity SEM 150 -nm Ni object pinhole 100 -nm Ni image pinholes object pinholes vector E-M field simulations 25 nm diffraction angle

The absence of astigmatism confirms accuracy 39 -nm isolated lines 0. 1 -NA ETS optic: lithography at LBNL 13. 5 nm wavelength = 0. 7, partial coherence DOF = ± 0. 5 µm EUV-2 D resist, 120 -nm thick Coded as 80 nm (1: 1) narrowed by exposure bias (x 1. 4) Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

0. 3 -NA MET: Modulation down to ~25 nm 45 nm 30 nm 1. 8 mm Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16 35 nm 25 nm Rohm and Haas photo-resist R. Brainard, et al.

Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

Direct visible / EUV comparisons lead to improved accuracy Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

We also cross-calibrate the different EUV techniques visible-light EUV PS/PDI EUV shearing • Inter-comparisons have also improved EUV testing methods. Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

Conclusions • High-quality EUV optics are available now largely due to the development of EUV lithography slope errors: roughness: micro-roughness: < 1 µrad ≤ 1– 2 Å • Ultra-high accuracy interferometry is required for fabrication and alignment: visible (LLNL) and EUV (LBNL). • We have extended our EUV measurement techniques to 0. 3 NA, ƒ/1. 67 – Achieved Wavefront quality of 0. 55 nm ( /24. 5). – Demonstrated interferometer accuracy of 1 Å, and below. Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

Extra slides follow. . . Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

Interferometry across the field of view image-plane field measurement [nm] Note: diameter ~ magnitude • Initial measurements showed a large spherical aberration 0. 8– 0. 9 nm • Field measurements are made at 9(x, y) 3(z) = 27 points, 135 interferograms, ~ 3 hours. Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

Using interferometry to optimize the MET alignment image-plane field measurement [nm] Note: diameter ~ magnitude • Reduced the aberration magnitude to 0. 55 nm = /24. 5 • Set astig. , coma, and sph. ab. to ~ /225 – /340 Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

Different kinds of interferometry are used Visible-light • Sommargren Interferometer (LLNL) EUV • Knife-edge (Foucault) testing • Lateral Shearing Interferometry (LSI) - cross-grating technique • Phase-shifting point-diffraction interferometry - ultra-high accuracy - invented at CXRO • Hartmann test - best for low-NA Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16

EUV deviation from a perfect, spherical wavefront = 13. 4 nm Kenneth Goldberg, KAGoldberg@lbl. gov, SPIE 2005, 5900– 16 astigmatism coma spherical trifoil 37 RMS 111 pm 47 pm 292 pm 703 pm