The study for image placement repeatability of EUV mask on the flat chuck

Transcription

1 The study for image placement repeatability of EUV mask on the flat chuck Shusuke Yoshitake, Hitoshi Sunaoshi, Shuichi Tamamushi, Soichiro Mitsui 1, Munehiro Ogasawara 1, Takeyuki Yamada 2, Tsutomu Shoki 2, Jörg Butschke 3, Mathias Irmscher 3, Michael Ferber 4, Jochen Bender 4, Dieter Adam 4, Klaus-Dieter Röth 4 NuFlare Technology, Inc. 1. Toshiba Corporation 2. HOYA Corporation 3. Ims chips 4. Vistec Semiconductor Systems GmbH

2 Outline Introduction Advantages of and challenges for EUV mask Error budget analysis of IP requirement Offline Blank Backside Topography Correction, OBBTC Strategy for mask writer development Experiment 1 Registration performance of EBM-5 Writing test of EBM-5 with OBBTC Experiment 2 Blank materials Measurement repeatability of LMS-IPRO2 Two types of vacuum chuck Discussion Dynamic measurement repeatability with plate loading/unloading Image placement repeatability on the flat chuck Conclusion

3 Advantage of and challenges for EUV mask Strong candidate for lithography of 32nm and beyond Resolution Advantages Data volume; can be relaxed because of larger k1 factor Image Placement (IP); comparable to optical mask EUV mask may require relaxed IP compared to optical mask with DPT. Challenges Particle control ~ eliminate particles of 3nm in production of 32nm EUV mask An EUV mask has to be held on Electro-Static Chuck (ESC) in EUV scanner. Mask flatness variation, Thickness variation, mask local slope, bending by chucking, etc.

4 Error budget analysis of IP requirement IP [3σ (nm)] ITRS roadmap 65nm hp 45nm hp Optical mask 32nm hp EUV mask 1. Expected to meet the spec of ITRS k2:.5, NA: ESCs were used in EB writer, metrology and scanner. 4. Expected to meet the spec of ITRS 25 ITRS Requirements IP (multipoint) Blank flatness Mask flatness error Mask local slope Mask deformation by chucking 3 Sub total (Specific error of EUV mask) [Unit; nm] EB writer Total IP error The error specific to EUV mask is too large to meet the IP requirement on 25 ITRS roadmap. The urgent challenge is to find the solutions to mitigate the EUV-Mask-specific error within the total IP error.

5 OBBTC *: EUVL Symposium Obtain the blank backside topography by flatness measurement tool Obtain local gradient from the blank backside topography Obtain local pattern shift from local gradient map Obtain the coefficients of 3rd order polynomials for EB system Add new coefficients to system parameters for each patterning job X th order polynomials fitting Plate 1 Y x 1 = a 1 +a 11 x+a 12 y+a 13 x 2 +a 14 xy+a 15 y 2 +a 16 x 3 +a 17 x 2 y+a 18 xy 2 +a 19 y 3 y 1 = b 1 +b 11 x+b 12 y+b 13 x 2 +b 14 xy+b 15 y 2 +b 16 x 3 +b 17 x 2 y+b 18 xy 2 +b 19 y X 1 = (a s +a 1 )+(a s1 +a 11 )x+(a s2 +a 12 )y+(a s3 +a 13 )x 2 +(a s4 +a 14 )xy+ (a s5 +a 15 )y 2 +(a s6 +a 16 )x 3 +(a s7 +a 17 )x 2 y+(a s8 +a 18 )xy 2 +(a s9 +a 19 )y 3 Y 1 = (b s +b 1 )+(b s1 +b 11 )x+(b s2 +b 12 )y+(b s3 +b 13 )x 2 +(b s4 +b 14 )xy+ (b s5 +b 15 )y 2 +(b s6 +b 16 )x 3 +(b s7 +b 17 )x 2 y+(b s8 +b 18 )xy 2 +(b s9 +b 19 )y 3 S: system parameter for IPRO grid

6 Strategy for mask writer development EUV is not the only solution for NGL. DPT and Hyper-NA immersion are also the candidates for 32 nm lithography. Early access to EB writer for EUV mask process development is important EB writer users are not well positioned to make a selection of EUV specific writer. Capability to write conventional optical mask and EUV mask without any hardware modification is a compromised solution for 45nm and 32nm mask development. No ESC for EB writer.

7 Experiment 1 EBM-5 was used to understand the correction accuracy. OBBTC proved its certain practical functionality in previous experiment using EBM-4*. To further study, use of EBM-5 with better IPA and IPR became necessary. Writing test of EBM-5 with OBBTC 6 test masks IPA and IPR of EBM-5 measured by IPRO Residuals error of OBBTC *: EUVL Symposium 25

8 Image placement of EBM-5 measured by IPRO Image placement accuracy Image placement repeatability

9 Plate 1 EB IPRO Comparison Image Placement Accuracy (nm) EB IPRO Comparison X Y X Y X Y 3 σ Max Min

10 Residual errors of OBBTC Image placement accuracy on IPRO [3σ (nm)] X Y Image placement accuracy on the assigned grid [3σ (nm)] Image placement differences [3σ (nm)] X Y X Y Plate Plate Plate Plate Plate Plate There are no significant error caused by OBBTC. Residual errors of IPA and IPR of EBM-5 with OBBTC were comparable to IPA and IPR without OBBTC. The writing grid of EBM-5 was successfully corrected to the assigned grid by OBBTC.

11 Experiment 2 Vacuum chuck was used for the verification that the system grid of the latest EB writer can be adjusted with sufficient accuracy to the grid on the flat chuck by OBBTC. Blank materials Flatness variation of test masks were intentionally selected. Measurement repeatability of LMS-IPRO2 Two types of vacuum chuck were prepared. Vacuum chuck with holes has better flatness. Vacuum chuck with pedestals has 5 times stronger chucking force.

12 Flatness variation of test blanks Plate 1: 1.59 μm Plate 2: μm Plate 3:.2 μm Plate 4:.13 μm Plate 5:.731 μm Plate 6:.639 μm Topography data and blank materials were prepared by HOYA.

13 IPR of EBM-5 by LMS-IPRO2 LMS-IPRO LMS-IPRO2

14 Flat chuck Chucking force; 1586 N Chucking force; 315 N mm square φ 4 mm, pitch 8mm Vacuum chuck with pedestals (VCP) Vacuum chuck with holes (VCH) Differences between previous test chuck and new test chuck Chucking surface flatness of VCH is better than VCP. Chucking area of VCP is five times larger than VCH. Chucking force of 315N ~ 145mm x 145mm with 15 kpa.

15 Dynamic measurement repeatability with plate loading/unloading Dynamic measurement repeatability [3σ (nm)] VCP case Flatness [μm] Dynamic measurement repeatability [3σ (nm)] Flatness [μm] VCH case Plate 1 Plate 2 Plate 3 Plate 5 Plate 7 Plate 8 Plate 1 Plate 2 Plate 3 Plate 4 Plate 5 Plate 6 X Y Flatness X Y Flatness Better measurement repeatability with VCP obtained than VCH. Dependency on blank backside topography also observed with VCP. In case of VCH, dependency on the blank backside topography was unclear. This might be due to insufficient chucking force of VCH. Chucking force is important to obtain sufficient measurement repeatability with the flat chuck.

16 Image placement repeatability with OBBTC on the flat chuck All plates; Plate 1 ~ 6 IPR [3 σ (nm)] VCH case X Y 18 x x x x x VCP case IPR [3 σ (nm)] X Y 17 x x Measured by LMS-IPRO2

17 Image placement repeatability with OBBTC 1x1 1x1 1x1 VCH Image placement repeatability [3σ (nm)] Flat; 3 and 4 Convex; 1 and 6 Concave; 2 and 5 X Y X Y X Y 18 x x x x x Measured by LMS-IPRO2 VCP Best case X: 22.87, Y:17.1

18 Discussion 1 Dynamic measurement repeatability with loading/unloading In case of the vacuum chuck with pedestals (VCP), better measurement repeatability on the flat chuck was obtained. Dependency on the blank backside topography also observed. Dependency of the blank backside topography is unclear in case of vacuum chuck has holes (VCH). This might be due to insufficient chucking force of VCH. Chucking force is important to obtain sufficient measurement repeatability with the flat chuck. Image placement repeatability with OBBTC Image placement repeatability and the blank backside topography are correlated. Concave plates are better than convex plates and flat plates to obtain IPR with VCH. Chucking surface has a concave shape. Therefore, the variation of plate deformation by chucking was smaller than other type of plates.

19 Discussion 2 Proposals for IP management of EUV mask Correction technique is applicable with sufficient accuracy. Mask distortion by ESC Data management of the topography of ESC for each scanner IP error related to blanks Offline blank backside topography correction, OBBTC The correction of blank surface topography and thickness variation was also possible. Important subject: measurement repeatability The conditions for sufficient measurement repeatability on the flat chuck could not be identified as yet. 3PS of LMS-IPRO and EB writer is repeatable. IPA and IPR on the vacuum chuck with higher chucking force will be tested in next step.

20 Conclusion OBBTC, developed for EUV mask application, proved its certain practical functionality with EBM-5. The writing grid was successfully corrected to the predicted grid according to the backside topography of each plate. The vacuum chuck for LMS-IPRO2 was developed to verify the image placement repeatability with chucking. Blank flatness and measurement repeatability are correlated under the condition of higher chucking force. Correlation between blank flatness and measurement repeatability was unclear when vacuum chuck with holes (VCH) was used. The repeatable grid is the key for the grid correction. IPA and IPR on the vacuum chuck with higher chucking force and better flatness will be tested in next step.

21 Thank you for your attention!