High Brightness EUV Light Source for Metrology

Transcription

1 High Brightness EUV Light Source for Metrology Sergey V. Zakharov, Peter Choi, Vasily S. Zakharov EPPRA sas NANO UV sas Panel Discussion: Actinic Defect Inspection Technology for EUV MasksM

2 3 Remaining Focus Areas EUVL Symposium, Tahoe 08 EUVL Symposium Steering Committee identified three remaining focus areas that the industry needs to work on to enable EUVL manufacturing insertion: Long-term source operation with 100 W at the IF and 5 megajoule per day Availability of defect-free masks, throughout a mask lifecycle, and the need to address critical mask infrastructure tool gaps, specifically in the defect inspection and defect review area Simultaneous resist resolution sensitivity and LER ( ntent=techreport&utm_campaign=q308) - light source for Litho and mask inspection critical - COPYRIGHT NANO UV

3 Actinic Mask Inspection Tool Current State of the Art Worldwide the AIT and the EUVM are the highest performing EUV microscopes currently available for EUV mask research. However, to support pilot line operation and EUVL transition into manufacturing, a commercial EUV aerial imaging tool will be required for patterned mask defect review. - Synchrotron Light Source necessary - COPYRIGHT NANO UV

4 Metrology Source a very different source compared with Litho sources illumination field size - much smaller ( mm) 2 power density on target - irradiance much higher etendue - much smaller ( ~ mm 2.sr ) source brightness - much higher sufficient throughput mask blank inspection < 2 hrs pattern mask inspection throughput of < hrs for full mask AIMS throughput of < hrs for full mask different optics compared with Litho scanner small field size - diffractive optics small NA on Mask to match projection optics - Source brighter than a synchrotron is needed - COPYRIGHT NANO UV

5 EUV Brightfield Metrology - Requirements Mask defect inspection - the source required - relative defect response > N photon statistics Consider a CCD array (n n) detector, pixel size A p, being used to image the area of the mask under inspection - magnification of imaging optics, m, hence area to detect a defect is now A i =A p /m 2, and the total illuminated patch area on mask observed is A=A i n 2 - total illumination time: t =t A M m 2 /n 2 A p illuminating irradiance required: A t A R D t n - then for defect size 10 nm, a (9μm) 2 pixel size, CCD array and full size (4 2 (26 33) mm 2 ) mask inspection: N A D 2 N > A N Magnification, m Patch area, A (um2) 5.06E E E-03 Illuminating flux density (ph/cm2) 5.47E E E+14 Na illuminating A 1.16E E E+10 Irradiance at mask needed, 10 shots exposure (ph/s cm2) 2.74E E E+17 Mask exposure time (min) 2.16E E E+01 > 4 M (K. Goldberg, Hawaii, 2008) N A D A i A *additional time for positioning and alignment needed in each exposure N M (reflectivity R 60%) COPYRIGHT NANO UV

6 Optimized EUV Efficiency of a Source Plasma self-absorption defines the limiting brightness of a single EUV source and required radiance The plasma parameters where EUV radiance is a maximum are not the same as that when the spectral efficiency is a maximum. The spectral efficiency is high SE=11-12% at plasma density ρ = g/cm 3. The optimum Conversion Efficiency@IF (CE IF ) of in-band radiation is maximal at target radius value smaller than R< 0.38mm and decreases like R -2 at larger one. - the Conversion Efficiency of a single source decreases if the in-band EUV output. increases (at the same operation frequency) EUV Radiance, MW/mm2 sr Spectral Efficiency (Peuv/Prad) tin Effective Depth (rho2*r), g2/cm3 g 2 5 tin Z* Scan R=0.04mm R=0.08mm R=0.16mm R=0.31mm R=0.625mm R=1.25mm R=2.5mm R=5mm R=0.04mm R=0.08mm R=0.16mm R=0.31mm R=0.625mm R=1.25mm R=2.5mm R=5mm COPYRIGHT NANO UV Effective Depth (rho2*r), g2/cm5 g 2 5

7 What is achievable with a single source? - LPP Critical density for Nd:YAG laser (ρ cr g/cm 3 ) is in the region of low spectral efficiency (SE). Example: to obtain <L EUV >=1W/mm 2 sr= L EUV τ f (τ R /C s ) corresponding to the inband power P EUV = <L EUV > 16πA s from conventional source the laser power required P l P EUV /η rad SE in optimum P l >2kW (η rad 0.5) operating with frequency f > 1kHz with R=63μm. The Spectral Efficiency (Peuv/Prad) tin Z* Scan R=0.04mm R=0.08mm R=0.16mm R=0.31mm R=0.625mm R=1.25mm R=2.5mm R=5mm Density (rho), g/cm3 usable power with etendue mm 2 sr is W, i.e CE IF %. For LPP source based on CO 2 -laser the critical density is 100 times less. Example: to obtain <L EUV >=1W/mm 2 sr the spectral efficiency may be higher, up to SE 11%, but lower L EUV resulting in laser power required P l >1kW operating with khz f R -4. CO 2 -laser EUV source requires an additional spectral purity filter and anti-fast-ion protection that reduces the effective CE. COPYRIGHT NANO UV

8 What is achievable with a single source? - DPP In conventional single DPP source of Z-pinch type the plasma is heated up and compressed by magnetic field energy and pressure (B 2 /8π ). From Bennet ratio ρr M i R B 2 /(8π T(Z+1)) I 2 [A] / R [cm] The collectable in-band EUV power for given etendue E is [mm 2 sr] I P EUV (W) I 2 [A] E f [Hz] The joule heating should provide enough power: I 2 / (π σ) f > P EUV 16 π 2 R 2 / (E SE) R [cm] < 0.13 SE 0.5 To have high P EUV and radiance <L EUV > P EUV /16π2 R 2 from a single source in the optimal regime of high EUV radiance the current is high I [A] ~ R 0.5 supplied during short time τ(s) R [cm]. Example: the radiance obtained <L EUV >=1W/mm 2 sr corresponds to the conventional source in-band power P EUV =16W operating with the current I =8 ka and f > 4 khz. This current should be supplied to a small size R<0.3mm plasma during 50ns. The power required increases P R 2 with plasma size. B COPYRIGHT NANO UV

9 EUV IF Power Limitation: Xenon plasma EUV emission prediction vs. observation EUV Radiance, MW/mm2 sr xenon R=0.04mm R=0.08mm R=0.16mm R=0.31mm R=0.625mm R=1.25mm R=2.5mm R=5mm Mass Depth (rho*r), g/cm2 Xenon plasma parameter scan with Z*-code showing the EUV radiance limitation Experimental observation of limitation of the EUV power at IF from xenon DPP source (M. Yoshioka et al. Alternative Lytho. Tech. Proc. of SPIE, vol () COPYRIGHT NANO UV

10 Emissivity, a.u. Total Emission of Xe XXI-Xe Xe XXIV ions from plasma with e-beame EUV Measurement Capillary discharge. VUV spectrograph data 0.1 ar) in o Wavelength (nm) o.4 o.3 o.2 pressure (mbar) Intensity (arb. units) 0.01 Xe 33 ev Total Xe 80eV + 2% 3keV Wavelength, nm Total EUV emission spectra of Xe XXI - XXIV ions from non-equilibrium plasma at 80 ev with 2% of fast electrons at 3 kev in comparison with emission spectrum of Xe XI ions from plasma at 33 ev (black). Electron density N e = /cm 3 COPYRIGHT NANO UV

11 Multiplexing - a solution for high power & brightness For small size source, the intensity of the inband emission is maximal in almost transparent plasma at ρr = ( ρr), near κωr ~ 1 The etendue from a single small size source is low enough E 1 =A s Ω << 1 mm 2 sr to be multiplexed. The EUV power of multiplexed N sources is P EUV as n i = Np n n Finally : P i EUV e EUV πr N ( ρ R R ~ To increase the power significantly we have to increase the plasma dimension R, or operation frequency f, or number of sources N. The EUV source power meeting the etendue requirements increases as N 1/2 2 ) E N Ω t f, Ω t f ; as EUV Radiance, MW/mm2 sr E /( N Ω) tin Z* Scan Mass Depth (rho*r), g/cm2 R=0.04mm R=0.08mm R=0.16mm R=0.31mm R=0.625mm R=1.25mm R=2.5mm R=5mm Decreasing the plasma size doesn t reduce the EUV radiance (if the plasma optical depth is kept constant) This allows efficient re packing of radiators from 1 into N separate smaller volumes without losses in EUV power - problem is the physical size of SoCoMo COPYRIGHT NANO UV

12 High Brightness EUV Source i-socomo Pilot production unit spec compact form factor mm spot size at up to 1 m distance photons/cm 2 4% BW inband EUV irradiance 3 khz continuous 3.3 kw average power consumption 1 Gshot lifetime to service in-build photon collection & projection plasma structure - PlasmaLens CYCLOPS BE-16 Discharge voltage EUV COPYRIGHT NANO UV

13 CYCLOPS BE-16 for Metrology Exceptional source brightness* and power** at 3 khz * irradiance greater than EUV photons/cm 2 /s measured at 64 cm from the source over a 5 mm 2 area, 13.5nm, 4%BW ** equivalent source power - more than 30 kw (2π sr, 13.5 nm, 4% BW) COPYRIGHT NANO UV

14 Source Characteristics Emission properties measured time averaged source diameter 0.6 mm FWHM 1.02 mm 2 spot size (1/e 2 ) emission angle source etendue EUV radiant brightness* 0.32 half angle 9.5 e-5 steradian 9.7 e-5 mm 2.sr 4.5 e18 photons/(mm 2.sr.s)/4% BW [1s average] 8.3 e22 photons/(mm 2.sr.s)/4% BW [1 pulse peak] * using average signal on SXUV5 diode of photons (91 ev) after 2 ML reflection - a very bright source for metrology - signal Source image through 400 μm pinhole - 1 min exposure, 6 E4 shots integrated image sensor at 108 cm from source Data1_B Gauss fit of Data1_B Data: Data1_B Model: Gauss Chi^2/DoF = R^2 = y ± xc ± w ± A ± dimension (mm) Guassian fit of emission profile recorded COPYRIGHT NANO UV

15 PlasmaLens Photodiode signal (mv) Photodiode signal (mv) Distance (mm) cm Gaussian 44cm Gaussian fit Distance (mm) Slit-scan spatial profile Optical properties Radial distance(mm) single Gaussian profile fitting to obtain radiation half width source diameter (1/e 2 spot size) 2.5 mm at 44 cm from exit of PCS; 69 cm from plasma source Half width Linear fit Half Half angle= angle= º Solide angle= sr -5 sr Axial distance from end of collimator (cm) COPYRIGHT NANO UV - very small etendue: < 10 4 mm 2.sr -

16 Wavefront Characteristics HASO X-EUV Shack-Hartmann wavefront sensor measurements * Movement of the focalization point in (µm) Source pointing stability Exposure duration = 60s Horizontal position Vertical position Exposure number Beam divergence half angle =0.18 Exposure time = 60s at 1kHz repition rate COPYRIGHT NANO UV EUV beam diameter = 9.1mm at distance between CCD and focal spot =1430mm Etendue =4.4e 5 mm 2 sr * with support of G. Dovillaire, E. Lavergne from Imagine Optic and P. Mercere, M. Idir from SOLEIL Synchrotron

17 HYDRA -12BE18 Design Specification - a EUV source for mask metrology photons/cm 2 /s/2% BW in band EUV irradiance 12x i SoCoMo units working at 5 khz each configurable pupil fill etendue ~ 10 2 mm 2.sr COPYRIGHT NANO UV

18 HYDRA -12BE18 System Performance - a 40 khz continuous source spatial multiplexing well behaved small cross talk in sequential operation plug and play needs to tune individual cell 10x i SoCoMo units commissioned so far GEN-II CYCLOPS cells 10 x 4 khz 40 khz COPYRIGHT NANO UV End-on view of 10 sources

19 Acknowledgements Collaborators Pontificia Universidad Catolica de Chile RRC Kurchatov Institute, Moscow, Russia Keldysh Institute of Applied Mathematics RAS, Moscow, Russia University College Dublin King s College London Sponsors EU & French Government ANR- EUVIL OSEO-ANVAR RAKIA EUV LITHO, Inc. COPYRIGHT NANO UV

20 Acknowledgements R&D & Product Engineering team Raul Aliaga-Rossel, Aldrice Bakouboula, Otman Benali, Philippe Bove, Michèle Cau, Yves Chemla, Grainne Duffy, Olivier Coulibaly, Sebastian Fant, Julien Kuhn, Blair Lebert, Ouassima Sarroukh, Luc Tantart, Clement Zaepffel COPYRIGHT NANO UV