Large-Area Interference Lithography Exposure Tool Development

Transcription

1 Large-Area Interference Lithography Exposure Tool Development John Burnett 1, Eric Benck 1 and James Jacob 2 1 Physical Measurements Laboratory, NIST, Gaithersburg, MD, USA 2 Actinix, Scotts Valley, CA 2011 International Symposium on Lithography Extensions, Miami, October 20-21, 2011

2 Maskless Hybrid IL Concept LONG COHERENCE LENGTH FAR-UV LASER Seeder PLMA NLO LOW COST NANO-SCALE ASICS DARPA GRATE

3 Project Goals Demonstrate feasibility of large-size interference lithography Demonstrate: 1) New 197 nm laser appropriate beam characteristics to enable die-field size patterns (~33 mm x 26 mm). The issues are power, spatial mode quality, band width, and stability. 2) Can deliver beams to wafer with phase properties to enable pattern fidelity. 3) Metrology concept to control interference pattern pitch and pattern registration. Concept based on moiré patterns. 4) Develop grating fabrication tool verify complete concept.

4 Mach-Zehnder design based on plate beam splitter with large-diameter beam (40 mm) Advantages: Large interference pattern possible (~33 Large DOF. 26 mm). COTS optics meet requirements (no special gratings). Challenges: IL Approach Interference from opposite sides of beam requires high spatial coherence. Large pattern requires long temporal coherence L c. Large L c poses coherent scattering issues (unwanted reflection patterns, coherent defect scattering). Interference pattern is ray mapping rather than imaging need to polish P-V (not RMS). Temporal Coherence: L c = 197 nm: = pm L c = 30 cm

5 Laser Concept Generating narrow-band, high-power, sub 200 nm light Use a stable, coherent, tunable CW IR fiber laser seeder Chop into two-nanosecond pulses at 1-4 MHz rep rate Amplify in large mode area fiber amps EO phase modulator to compensate chirp pulses Frequency upshift IR to UV using efficient non-linear optical processes

6 Infrared Front End of Light Source Pulse timing, chirp compensation and drive electronics BP filter Diode Pump <10 W CW Fiber Laser 1055 or 1550 nm 25 mw, 5 khz LW Modulator 1-2 ns PW 1-4 MHz PRF Isolator Phase compensator Fiber pre-amp Fiber power amp 2.5 W Avg Power 1 MHz PRF, 2 ns PW 400 MHz BW [Approved for Public Release, Distribution Unlimited]

7 197 nm System Fiber Laser 1055 nm Freq Doubler Freq Tripler Freq Mixer Freq Mixer Fiber Laser 1550 nm Freq Doubler 197 nm 250 mw 0.75 ns 0.13 pm BW High conversion efficiency relaxes dependence on fiber front ends to produce high peak power, which in turn reduces the amount of SPM needed to be compensated [Approved for Public Release, Distribution Unlimited]

8 Beam Shaping Must convert small Gaussian laser beam to uniform collimated beam. Keplerian-type anamorphic telescope Gaussian input beam collimated beam with planar wavefront first lens redistributes rays to transform the intensity profile second lens corrects the wavefront distortion due to the first lens Gaussian input beam distributions (before homogenizing lens) Ray trace model for 197 nm laser homogenizing lens (asphere) 40 mm Flat top output distribution (before collimating lens) Actinix 197 nm laser Beam diam. at 1 st lens: ~1 mm Divergence: ~0.7 mrad (FWHM -1/2 angle) Collimating/wave flattening Aspheric group

9 Beam Shaper Aspheric Surfaces Gaussian input beam collimated beam with planar wavefront Issues Geometric Optics 2 convex aspheric surfaces can be used to exactly transform a Gaussian beam to an arbitrary output profile, e.g. top hat. Top hat profile gives substantial diffraction due to edges affects beam parallelism and uniformity on propagation < 1m. Need to roll off output profile, e.g., Fermi-Dirac. R 0 governs range, governs roll off Details depends on precise characteristics of laser to determined. Beam shaper ray-mapper - light from each source point tracks through 1 point on asphere surface must figure asphere to P-V specifications! For interferometer, figure errors result in fringe positioning errors. For control of fringe positions: Angle tolerance ~ ~0.05 mrad, P-V tolerance ~ 50 nm.

10 Interferometer - Modeling Actinix 197 nm solid-state laser fine frequency control - pitch control beam shaping optics HR mirror - pitch control 70 nm pitch beam splitter compensator 40 mm deformable mirror coupling prism CCD plate beam splitter imaging optics wafer/ metrology grating HR mirror - pitch control Realistic ray trace model Input beam from beam shaper. Trace beam to wafer plane - gives 35 nm HP interference pattern. For L c = 30 cm in model: no significant loss in image contrast at edge of field (16.5 mm). Started modeling use of adaptive optics to correct effects of aberrations on pattern. Interference pattern at edge of field 33 mm 26 mm How do you insure pattern has correct pitch, orientation, position? Real-time pattern metrology. Interference pattern at center of field

11 Moiré Interferometry Concept Input Beam 1 Input Beam 2 2 beams with half angle pattern P= /2 sin Superpose on grating with spacing d Condition that 1 st orders diffract normally: d= /sin or d = 2p Two waves have constant phase relation 1, -1 2, +1 d = grating pitch If conditions are not precisely met, a moiré beat pattern is produced. Modulation envelope with period p mod = -p(p/ p). Basis for metrology scheme to characterize pitch deviations from reference grating and correct with feedback control.

12 Moiré Interferometry Projection of interference pattern on reference grating with line spacing = 2 gives ±1 diffraction orders in vertical direction for both beams. Image beams on CCD. If interference pattern/grating lines not commensurate moiré pattern. pitch Moiré pattern gives deviation from perfect overlay of interference pattern on reference grating. Eliminate moiré nulls with stage rotation, translation, control, and adaptive optics. Can be used in feedback mode to correct overlay error. Actinix 197 nm solid-state laser fine frequency control - pitch control beam shaping optics HR mirror - pitch control beam splitter compensator 40 mm deformable mirror coupling prism CCD plate beam splitter imaging optics wafer/ metrology grating reference grating w/ 2 interference pitch at wafer plane HR mirror - pitch control

13 Moiré Interferometry - Modeling Quantitative simulation of optical effect projection of 197 nm interference pattern (35 nm HP) on reference grating =8.7 rad, =3 pm, lens aberration =8.7 rad, =3 pm, no lens aberration =0 rad, =3 pm, Remove wavefront distortion w/ adaptive optics Reduce vertical component of moiré beats w/ stage rotation =0.87 rad, =0.1 pm, =0.87 rad, =0.1 pm, stage translation=35 nm =0 rad, =0 pm Reduce horizontal component of moiré beats w/ adjustments Establish correct phase with stage translation Maintain no moiré nulls with feedback control Complete elimination of moiré nulls guarantees registration across field!

14 Moiré Interferometry Measurements 266 nm MZ Interferometer with grating at interference plane. mirror 1 spatial filter/beam expander 50 mm 266 nm laser 2400 gr/mm grating (416 nm pitch) Beams projected at 39.7 from normal. imaging optics CCD beam splitter mirror 2 Images of interference pattern projected on grating for various mis-registrations. 0, 0 0, 0 0, reduced 0, reduced 0, 0 Moiré null region 1 cm interference pattern (104 nm half pitch) commensurate with grating lines. Verifies basic metrology concept at 266 nm

15 Summary Program underway to explore viability of full-field Interference Lithography. Actinix/Sandia building the laser which can be used for IL and inspection projects. Have modeled some of the key optical issues coherence, beam shaping. Have devised viable approach to metrology to ensure registration. Have modeled the metrology concept and demonstrated the principle at 266 nm. Plan and Prospects NIST continue building metrology system; Actinix completing laser breadboard. Incorporation into lithography demonstration tool pending funds and partners. We strongly believe the concept is viable for low volume, low cost nano-fabrication. Acknowledgments We gratefully acknowledge Michael Fritze for pioneering the hybrid lithography concept and supporting of our efforts, John Hoffnagle for key contributions with the beam shaping and coherence length modeling, and Darrell Armstrong for his work on the fiber laser development. This work is supported in part by DARPA contract W91CRB-10-C-0080.