December 15, 2014 / Vol. 39, No. 24 / OPTICS LETTERS
6969
Fractional Talbot lithography with extreme ultraviolet light Hyun-su Kim,1,4 Wei Li,3 Serhiy Danylyuk,2 William S. Brocklesby,4 Mario C. Marconi,3 and Larissa Juschkin1,* 1
2
Chair for the Experimental Physics of EUV, RWTH Aachen University and JARA—Fundamentals of Future Information Technology, Steinbachstrasse 15, 52074 Aachen, Germany
Chair for the Technology of Optical Systems, RWTH Aachen University and JARA—Fundamentals of Future Information Technology, Steinbachstrasse 15, 52074 Aachen, Germany 3
Engineering Research Center for Extreme Ultraviolet Science and Technology, and Electrical and Computer Engineering Department, Colorado State University, Fort Collins, Colorado 80523, USA 4
Optoelectronics Research Center, University of Southampton, Southampton, SO17 1BJ, UK *Corresponding author: larissa.juschkin@rwth‑aachen.de Received October 13, 2014; accepted November 5, 2014; posted November 14, 2014 (Doc. ID 224904); published December 15, 2014
Fractional Talbot effect leads to the possibility to implement patterning of structures with smaller periods than the master mask. This is particularly attractive when using short wavelength illumination in the extreme ultraviolet because of attainable resolution in the sub-100-nm range. In this Letter, we demonstrate the Talbot lithography with the fractional Talbot effect under coherent illumination generated with a capillary discharge Ne-like Ar extreme ultraviolet laser. Various spatial frequency multiplications up to 5x are achieved using a parent grating. This technique allows a fabrication of nanostructures with high-resolution patterns, which is of high interest in many applications such as the manufacturing of plasmonic surfaces and photonic devices. © 2014 Optical Society of America OCIS codes: (050.1950) Diffraction gratings; (140.7240) UV, EUV, and X-ray lasers; (220.3740) Lithography; (220.4000) Microstructure fabrication; (220.4241) Nanostructure fabrication. http://dx.doi.org/10.1364/OL.39.006969
Extreme ultraviolet (EUV, 5–50 nm wavelength range) interference lithography (IL) [1] is a powerful method of fabricating sub-micron structures for many applications such as semiconductor devices [2], plasmonics [3], photonic crystals [4], quantum dots [5], etc. Talbot lithography is a mask-based IL technique using the Talbot effect that is capable to produce replicas of a periodic mask at specific positions defined by the Talbot distance, zT 2d2 ∕λ;
(1)
where d is the period of the structure [6]. Talbot IL under extreme ultraviolet illumination has been used for selfimaging of complex structures [7], high-resolution patterning of large areas [8,9], defect-tolerant printing under coherent illumination [10,11], and the patterning with spatial frequency multiplication in the intensity distribution using spectrally broadband radiation sources with the achromatic Talbot effect [12–14]. However, the spatial frequency multiplication or magnification, M sf , was limited to a factor of 2 in the technique based on achromatic Talbot effect. To achieve smaller structures in Talbot lithography, the high resolution masks fabricated with smaller periods are required. This fact imposes a limitation toward the printing of smaller feature sizes because of the challenge of producing large area, high quality, and small period masks compatible with EUV illumination [15]. To overcome this obstacle, we investigated a use of the fractional Talbot effect in Talbot lithography focusing on finding the alternative scaling of structures. Fractional Talbot effect describes the interference pattern not only at the integer multiples of the Talbot distance, n · zT , but also at other intermediate planes located at fractional Talbot distance. Particularly clear line and space patterns 0146-9592/14/246969-04$15.00/0
can be found at a rational number of the half Talbot distance, z p∕q · zT ∕2;
(2)
where p and q are coprime numbers [16]. The effect leads to fractional interference patterns from the parent mask for various M sf . While various imaging applications and analysis are reported on the fractional Talbot effect [17– 19], however, in the structuring, the fractional Talbot lithography was reported only with visible light and hard x-rays [20]. Herein, we demonstrate the fractional Talbot lithography in the EUV utilizing a compact EUV laser source, resulting in various M sf up to M sf 5 from one parent mask. The fractional Talbot pattern can be analytically expressed using the angular frequency representation [21]. The binary transmittance, E 0 x 1 in the slits and E 0 x 0 at the absorber, is considered as an initial intensity distribution at the mask plane z 0, where x is the axis of initial intensity distribution of grating and z is the axis of propagation of the plane wave illuminating the mask. In the paraxial limit, the angular frequency representation is identical with the framework of Fourier optics. The optical field, Ex; z is described with the inˆ x ; kz ; z verse Fourier transform of a propagator, Hk expi · kz · z in reciprocal space, where the longitudinal wavenumber is kz k2 − k2x 1∕2 , with the wavenumber, k 2π∕λ, thus Ex; z E 0 x; z 0 Hx; z. The intensity profile as a function of distance is illustrated in Fig. 1(a). The fractional Talbot images having positive integer M sf values 1x; 2x… are formed at z positions equal to fractional Talbot distances according to the following relations [also see Figs. 1(a) and 1(b)]: © 2014 Optical Society of America
6970
OPTICS LETTERS / Vol. 39, No. 24 / December 15, 2014
Fig. 1. Talbot pattern behind the grating with narrow slits (500 nm per 3 μm pitch) (a) as a function of distance from a grating, and (b) the intensity profiles at several distances indicating M sf —note that the cross-section data are averaged with 10 pixels. Fig. 2. (a) Profiles of the Talbot pattern in Fig. 1(a) along the radiation propagation axis at various x-positions showing the M sf and (b) the DOF for different M sf . Note that the number of DOF values measured for desired M sf equals M sf , and the values are slightly different. The DOF values in (b) are minimum one among those values.
1x: at z 1∕1 · n · zT ∕2; 0 n 1; 2… 2x: at z 1∕2 · n · zT ∕2 3x: at z 1∕3 · n · zT ∕2; 2∕3 · n · zT ∕2 4x: at z 1∕4 · n · zT ∕2; 3∕4 · n · zT ∕2 5x: at z 1∕5 · n · zT ∕2; 2∕5 · n · zT ∕2; 3∕5 · n · zT ∕2; 4∕5 · n · zT ∕2:
(3)
The distance between the mask and the sample has to be precisely controlled, within the depth of field (DOF) of the pattern to achieve the desired M sf . The DOF for the desired M sf is given approximately by 1∕2 · zMsf −1 − zMsf 1 ∕2 ≈ p2 ∕2λM 2sf − 1;
(4)
where zMsf 1∕M sf · zT ∕2 [20]. The DOF can be estimated from the simulation by measuring the peak intensity distribution of the Talbot pattern or Talbot carpet along the radiation propagation direction [22]. Figure 2(a) shows profiles of the Talbot image along the radiation propagation axis for several x-positions of Fig. 1(a). The profiles indicate the intensity range of the pattern for the desired M sf . The DOF corresponds to the full width halfmaximum (FWHM) of each peak for the respective M sf [see the red lines in Fig. 2(a)]. Figure 2(b) shows values of DOF at different M sf from 1x to 5x. As expected, for the higher M sf , the pattern yields the shorter DOF, which makes the distance control more critical. A capillary discharge Ne-like Ar laser at λ 46.9 nm was used for exposures in experiments. This laser provides high spatial and temporal coherence, which are required for this work. The spectral bandwidth was approximately Δλ∕λ 3.5 × 10−5 [23], and spatial coherence radius was 500 μm at the mask plane [24,25]. The
exposure dose at the wafer plane in the experiments was 0.1–0.3 mJ∕cm2 ∕pulse [11]. The transmission grating (or Talbot mask) was fabricated using focused ion beam (FIB) milling. A Tungsten probe tip was used to generate the W ion beam. The beam current was 80 pA and the dwell time was 1 μs during the process. The grating structure was defined on a Si3 N4 (50 nm thickness) membrane deposited with Au (130 nm thickness) that was fully milled through to make 20 slits and to achieve free standing array structures as shown in Fig. 3(b). In total, 20 pairs of the absorbing and opening lines were created with a 3 μm pitch over a total area of 60 μm × 50 μm. The number of pairs was enough to demonstrate the fractional Talbot effect, even though the line width roughness and the out-of-plane bending of the lines impacted the quality of results. The length of the lines was 50 μm and the width of the slits was ∼500 nm on average [see Fig. 3(a)–3(c)]. The grating was illuminated by the EUV laser to form the fractional Talbot pattern. A positive-tone photoresist [polymethyl methacrylate (PMMA)] with an approximate thickness 80 nm was deposited by spin coating on a Si wafer that was placed behind the mask at the corresponding fractional Talbot distances to achieve the desired M sf . The exposure dose was set at 100 shots (around 20 mJ∕cm2 ) through the experiment. After the exposure, the photoresist was developed with MIBK (Methyl isobutyl ketone): IPA 1∶3 for 45 s and cleaned with IPA for 45 s.
December 15, 2014 / Vol. 39, No. 24 / OPTICS LETTERS
Fig. 3. (a)–(c) Free standing transmission grating of narrow slit array formed on the Au deposited Si3 N4 window. Zoomin views (b) at the end of a slit and (c) in the middle.
The experimental results corresponding to the prints at the different fractional Talbot distances are summarized in Fig. 4, which are image data obtained using the atomic force microscope (AFM), showing the fractional Talbot lithography results of M sf 1 (a) M sf 2, (b) M sf 3, (c) M sf 5, and (d) with corresponding pitches of 3, 1.5, 1 μm, and 600 nm, respectively. In Fig. 4(a), the narrow slit on the mask is transferred into the wafer because of replication of the Talbot effect; Fig. 4(b) shows doubling of M sf . In Fig. 4(c), a small destructive interference at the peaks can be noticed because the wafer position was not exactly at z zMsf 3 , and in Fig. 4(d) line-edge roughness and out-of-plane bending of the parent mask
Fig. 4. AFM measurements of the lithography results showing (a) M sf 1, (b) M sf 2, (c) M sf 3, and (d) M sf 5 from the parent mask of 3 μm pitch. Overlayed red curves are corresponding AFM cross sections. The height of the PMMA structures (a)–(d) was ∼80 nm.
6971
Fig. 5. Intensity change in simulation (square blue) and developed depth of PMMA as a lithographic result (red circle) as a function of M sf .
start to degrade a quality of the pattern. However, the obtained results clearly show the expected multiplication factor at the frequency corresponding to the fractional Talbot effect. The widths of developed or removed parts were approximately (a) ∼570 nm, (b) ∼320 nm, (c) ∼310 nm, and (d) ∼210 nm at FWHM of the height, while the corresponding simulated values were respectively 670, 840, 780, and 840 nm at FWHM of intensity. According to the simulation data, the width is supposed to be comparable to the slit width of parent mask. The difference between the calculated values and the experiment data can be explained by variations in the illumination dose proper of the pulsed nature of the EUV laser and to the nonlinear response of the photoresist. The EUV transmission at a 46.9 nm wavelength is reduced approximately by half every 10 nm thickness of PMMA [26]. It also should be taken into account that peak intensity is changing as a function of M sf . The advantage of obtaining high M sf with fractional Talbot lithography is opposed by the fact that intensity reduces with increasing M sf as shown in Fig. 5. The intensity values from simulations (blue squares) and developed depth of the photoresist from experiments (red circles) are both decreasing for higher M sf , as expected. The fractional patterning with spatial frequency multiplication will be useful for applications that require large area and high-resolution patterning, relaxing the fabrication constrains for the master mask. It is also convenient and efficient when various sizes have to be printed using one mask in EUV patterning. In conclusion, we have demonstrated a fractional Talbot lithography experimentally using highly coherent EUV illumination at a 46.9 nm wavelength generated with the capillary discharge Ne-like Ar laser. Spatial frequency multiplications up to 5x were achieved with pitches of 3 μm, 1.5 μm, 1 μm, and 600 nm, respectively, from the parent grating of 3 μm pitch. The grating mask is highly transmissive with a free standing structure fabricated by FIB milling. However, the line width roughness and the out-of-plane bending of the lines must be improved for better quality of the printing over large areas. We envision that a grating with a smaller pitch size could be used for further reduction of period and feature size, and the
6972
OPTICS LETTERS / Vol. 39, No. 24 / December 15, 2014
grating with narrower slits could be used to get higher M sf values, more than M sf 5x. In addition, the utilization of recently developed EUV lasers in the vicinity of 13 nm can decrease further the feature size achievable with this patterning approach [27]. This work is performed within the EU FP7 Erasmus Mundus Joint Doctorate Programme EXTATIC under framework partnership agreement FPA-2012-0033. The authors also would like to thank Elmar Neumann from the Helmholtz Nanoelectronic Facility at the Forschungszentrum Juelich for help with FIB milling. References 1. H. H. Solak, J. Phys. D 39, R171 (2006). 2. The International Technology Roadmap for Semiconductors, http://www.itrs.net (2013). 3. E. Ozbay, Science 311, 189 (2006). 4. T. F. Krauss, R. M. De La Rue, and S. Brand, Nature 383, 699 (1996). 5. M. Reed, J. Randall, R. Aggarwal, R. Matyi, T. Moore, and A. Wetsel, Phys. Rev. Lett. 60, 535 (1988). 6. P. Latimer and R. F. Crouse, Appl. Opt. 31, 80 (1992). 7. A. Isoyan, F. Jiang, Y. C. Cheng, F. Cerrina, P. Wachulak, L. Urbanski, J. Rocca, C. Menoni, and M. Marconi, J. Vac. Sci. Technol. B 27, 2931 (2009). 8. L. Wang, H. H. Solak, and Y. Ekinci, Nanotechnology 23, 305303 (2005). 9. H. H. Solak, C. Dais, and F. Clube, Opt. Express 19, 10686 (2011). 10. L. Urbanski, A. Isoyan, A. Stein, J. J. Rocca, C. S. Menoni, and M. C. Marconi, Opt. Lett. 37, 3633 (2012). 11. W. Li, V. M. Esquiroz, L. Urbanski, D. Patel, C. S. Menoni, M. C. Marconi, A. Stein, W. Chao, and E. H. Anderson, J. Vac. Sci. Technol. B 31, 06F604 (2013).
12. N. Guerineau, E. D. Mambro, and J. Primot, Opt. Express 11, 3310 (2003). 13. H. H. Solak and Y. Ekinci, J. Vac. Sci. Technol. B 23, 2705 (2005). 14. S. Danylyuk, P. Loosen, K. Bergmann, H.-S. Kim, and L. Juschkin, J. Micro-Nanolith. MEM. 12, 033002 (2013). 15. S. Brose, S. Danylyuk, L. Juschkin, C. Dittberner, K. Bergmann, J. Moers, G. Panaitov, St. Trellenkamp, P. Loosen, and D. Grützmacher, Thin Solid Films 520, 5080 (2012). 16. M. V. Berry and S. Klein, J. Mod. Opt. 43, 2139 (1996). 17. J. Wen, Y. Zhang, and M. Xiao, Adv. Opt. Photon. 5, 83 (2013). 18. J. T. Winthrop and C. R. Worthington, J. Opt. Soc. of Am. 55, 373 (1965). 19. V. Arrizón and J. Ojeda-Castañeda, Appl. Opt. 33, 5925 (1994) 20. D. D. Flanders, A. M. Hawryluk, and H. I. Smith, J. Vac. Sci. Technol. 16, 1949 (1979). 21. J. W. Goodman, Introduction to Fourier Optics, 2nd ed. (McGraw-Hill, 1996). 22. W. B. Case, M. Tomandl, S. Deachapunya, and M. Arndt, Opt. Express 17, 20966 (2009). 23. L. Urbanski, M. C. Marconi, L. M. Meng, M. Berrill, O. Guilbaud, A. Klisnick, and J. J. Rocca, Phys. Rev. A 85, 033837 (2012). 24. Y. Liu, M. Seminario, F. G. Tomasel, C. Chang, J. J. Rocca, and D. T. Attwood, Phys. Rev. A 63, 033802 (2001). 25. M. C. Marconi, J. L. A. Chilla, C. H. Moreno, B. R. Benware, and J. J. Rocca, Phys. Rev. Lett. 79, 2799 (1997). 26. B. L. Henke, E. M. Gullikson, and J. C. Davis, At. Data Nucl. Data Tables 54, 181 (1993). 27. D. Alessi, D. H. Martz, Y. Wang, M. Berrill, B. M. Luther, and J. J. Rocca, Opt. Lett. 35, 414 (2010).
6969
Fractional Talbot lithography with extreme ultraviolet light Hyun-su Kim,1,4 Wei Li,3 Serhiy Danylyuk,2 William S. Brocklesby,4 Mario C. Marconi,3 and Larissa Juschkin1,* 1
2
Chair for the Experimental Physics of EUV, RWTH Aachen University and JARA—Fundamentals of Future Information Technology, Steinbachstrasse 15, 52074 Aachen, Germany
Chair for the Technology of Optical Systems, RWTH Aachen University and JARA—Fundamentals of Future Information Technology, Steinbachstrasse 15, 52074 Aachen, Germany 3
Engineering Research Center for Extreme Ultraviolet Science and Technology, and Electrical and Computer Engineering Department, Colorado State University, Fort Collins, Colorado 80523, USA 4
Optoelectronics Research Center, University of Southampton, Southampton, SO17 1BJ, UK *Corresponding author: larissa.juschkin@rwth‑aachen.de Received October 13, 2014; accepted November 5, 2014; posted November 14, 2014 (Doc. ID 224904); published December 15, 2014
Fractional Talbot effect leads to the possibility to implement patterning of structures with smaller periods than the master mask. This is particularly attractive when using short wavelength illumination in the extreme ultraviolet because of attainable resolution in the sub-100-nm range. In this Letter, we demonstrate the Talbot lithography with the fractional Talbot effect under coherent illumination generated with a capillary discharge Ne-like Ar extreme ultraviolet laser. Various spatial frequency multiplications up to 5x are achieved using a parent grating. This technique allows a fabrication of nanostructures with high-resolution patterns, which is of high interest in many applications such as the manufacturing of plasmonic surfaces and photonic devices. © 2014 Optical Society of America OCIS codes: (050.1950) Diffraction gratings; (140.7240) UV, EUV, and X-ray lasers; (220.3740) Lithography; (220.4000) Microstructure fabrication; (220.4241) Nanostructure fabrication. http://dx.doi.org/10.1364/OL.39.006969
Extreme ultraviolet (EUV, 5–50 nm wavelength range) interference lithography (IL) [1] is a powerful method of fabricating sub-micron structures for many applications such as semiconductor devices [2], plasmonics [3], photonic crystals [4], quantum dots [5], etc. Talbot lithography is a mask-based IL technique using the Talbot effect that is capable to produce replicas of a periodic mask at specific positions defined by the Talbot distance, zT 2d2 ∕λ;
(1)
where d is the period of the structure [6]. Talbot IL under extreme ultraviolet illumination has been used for selfimaging of complex structures [7], high-resolution patterning of large areas [8,9], defect-tolerant printing under coherent illumination [10,11], and the patterning with spatial frequency multiplication in the intensity distribution using spectrally broadband radiation sources with the achromatic Talbot effect [12–14]. However, the spatial frequency multiplication or magnification, M sf , was limited to a factor of 2 in the technique based on achromatic Talbot effect. To achieve smaller structures in Talbot lithography, the high resolution masks fabricated with smaller periods are required. This fact imposes a limitation toward the printing of smaller feature sizes because of the challenge of producing large area, high quality, and small period masks compatible with EUV illumination [15]. To overcome this obstacle, we investigated a use of the fractional Talbot effect in Talbot lithography focusing on finding the alternative scaling of structures. Fractional Talbot effect describes the interference pattern not only at the integer multiples of the Talbot distance, n · zT , but also at other intermediate planes located at fractional Talbot distance. Particularly clear line and space patterns 0146-9592/14/246969-04$15.00/0
can be found at a rational number of the half Talbot distance, z p∕q · zT ∕2;
(2)
where p and q are coprime numbers [16]. The effect leads to fractional interference patterns from the parent mask for various M sf . While various imaging applications and analysis are reported on the fractional Talbot effect [17– 19], however, in the structuring, the fractional Talbot lithography was reported only with visible light and hard x-rays [20]. Herein, we demonstrate the fractional Talbot lithography in the EUV utilizing a compact EUV laser source, resulting in various M sf up to M sf 5 from one parent mask. The fractional Talbot pattern can be analytically expressed using the angular frequency representation [21]. The binary transmittance, E 0 x 1 in the slits and E 0 x 0 at the absorber, is considered as an initial intensity distribution at the mask plane z 0, where x is the axis of initial intensity distribution of grating and z is the axis of propagation of the plane wave illuminating the mask. In the paraxial limit, the angular frequency representation is identical with the framework of Fourier optics. The optical field, Ex; z is described with the inˆ x ; kz ; z verse Fourier transform of a propagator, Hk expi · kz · z in reciprocal space, where the longitudinal wavenumber is kz k2 − k2x 1∕2 , with the wavenumber, k 2π∕λ, thus Ex; z E 0 x; z 0 Hx; z. The intensity profile as a function of distance is illustrated in Fig. 1(a). The fractional Talbot images having positive integer M sf values 1x; 2x… are formed at z positions equal to fractional Talbot distances according to the following relations [also see Figs. 1(a) and 1(b)]: © 2014 Optical Society of America
6970
OPTICS LETTERS / Vol. 39, No. 24 / December 15, 2014
Fig. 1. Talbot pattern behind the grating with narrow slits (500 nm per 3 μm pitch) (a) as a function of distance from a grating, and (b) the intensity profiles at several distances indicating M sf —note that the cross-section data are averaged with 10 pixels. Fig. 2. (a) Profiles of the Talbot pattern in Fig. 1(a) along the radiation propagation axis at various x-positions showing the M sf and (b) the DOF for different M sf . Note that the number of DOF values measured for desired M sf equals M sf , and the values are slightly different. The DOF values in (b) are minimum one among those values.
1x: at z 1∕1 · n · zT ∕2; 0 n 1; 2… 2x: at z 1∕2 · n · zT ∕2 3x: at z 1∕3 · n · zT ∕2; 2∕3 · n · zT ∕2 4x: at z 1∕4 · n · zT ∕2; 3∕4 · n · zT ∕2 5x: at z 1∕5 · n · zT ∕2; 2∕5 · n · zT ∕2; 3∕5 · n · zT ∕2; 4∕5 · n · zT ∕2:
(3)
The distance between the mask and the sample has to be precisely controlled, within the depth of field (DOF) of the pattern to achieve the desired M sf . The DOF for the desired M sf is given approximately by 1∕2 · zMsf −1 − zMsf 1 ∕2 ≈ p2 ∕2λM 2sf − 1;
(4)
where zMsf 1∕M sf · zT ∕2 [20]. The DOF can be estimated from the simulation by measuring the peak intensity distribution of the Talbot pattern or Talbot carpet along the radiation propagation direction [22]. Figure 2(a) shows profiles of the Talbot image along the radiation propagation axis for several x-positions of Fig. 1(a). The profiles indicate the intensity range of the pattern for the desired M sf . The DOF corresponds to the full width halfmaximum (FWHM) of each peak for the respective M sf [see the red lines in Fig. 2(a)]. Figure 2(b) shows values of DOF at different M sf from 1x to 5x. As expected, for the higher M sf , the pattern yields the shorter DOF, which makes the distance control more critical. A capillary discharge Ne-like Ar laser at λ 46.9 nm was used for exposures in experiments. This laser provides high spatial and temporal coherence, which are required for this work. The spectral bandwidth was approximately Δλ∕λ 3.5 × 10−5 [23], and spatial coherence radius was 500 μm at the mask plane [24,25]. The
exposure dose at the wafer plane in the experiments was 0.1–0.3 mJ∕cm2 ∕pulse [11]. The transmission grating (or Talbot mask) was fabricated using focused ion beam (FIB) milling. A Tungsten probe tip was used to generate the W ion beam. The beam current was 80 pA and the dwell time was 1 μs during the process. The grating structure was defined on a Si3 N4 (50 nm thickness) membrane deposited with Au (130 nm thickness) that was fully milled through to make 20 slits and to achieve free standing array structures as shown in Fig. 3(b). In total, 20 pairs of the absorbing and opening lines were created with a 3 μm pitch over a total area of 60 μm × 50 μm. The number of pairs was enough to demonstrate the fractional Talbot effect, even though the line width roughness and the out-of-plane bending of the lines impacted the quality of results. The length of the lines was 50 μm and the width of the slits was ∼500 nm on average [see Fig. 3(a)–3(c)]. The grating was illuminated by the EUV laser to form the fractional Talbot pattern. A positive-tone photoresist [polymethyl methacrylate (PMMA)] with an approximate thickness 80 nm was deposited by spin coating on a Si wafer that was placed behind the mask at the corresponding fractional Talbot distances to achieve the desired M sf . The exposure dose was set at 100 shots (around 20 mJ∕cm2 ) through the experiment. After the exposure, the photoresist was developed with MIBK (Methyl isobutyl ketone): IPA 1∶3 for 45 s and cleaned with IPA for 45 s.
December 15, 2014 / Vol. 39, No. 24 / OPTICS LETTERS
Fig. 3. (a)–(c) Free standing transmission grating of narrow slit array formed on the Au deposited Si3 N4 window. Zoomin views (b) at the end of a slit and (c) in the middle.
The experimental results corresponding to the prints at the different fractional Talbot distances are summarized in Fig. 4, which are image data obtained using the atomic force microscope (AFM), showing the fractional Talbot lithography results of M sf 1 (a) M sf 2, (b) M sf 3, (c) M sf 5, and (d) with corresponding pitches of 3, 1.5, 1 μm, and 600 nm, respectively. In Fig. 4(a), the narrow slit on the mask is transferred into the wafer because of replication of the Talbot effect; Fig. 4(b) shows doubling of M sf . In Fig. 4(c), a small destructive interference at the peaks can be noticed because the wafer position was not exactly at z zMsf 3 , and in Fig. 4(d) line-edge roughness and out-of-plane bending of the parent mask
Fig. 4. AFM measurements of the lithography results showing (a) M sf 1, (b) M sf 2, (c) M sf 3, and (d) M sf 5 from the parent mask of 3 μm pitch. Overlayed red curves are corresponding AFM cross sections. The height of the PMMA structures (a)–(d) was ∼80 nm.
6971
Fig. 5. Intensity change in simulation (square blue) and developed depth of PMMA as a lithographic result (red circle) as a function of M sf .
start to degrade a quality of the pattern. However, the obtained results clearly show the expected multiplication factor at the frequency corresponding to the fractional Talbot effect. The widths of developed or removed parts were approximately (a) ∼570 nm, (b) ∼320 nm, (c) ∼310 nm, and (d) ∼210 nm at FWHM of the height, while the corresponding simulated values were respectively 670, 840, 780, and 840 nm at FWHM of intensity. According to the simulation data, the width is supposed to be comparable to the slit width of parent mask. The difference between the calculated values and the experiment data can be explained by variations in the illumination dose proper of the pulsed nature of the EUV laser and to the nonlinear response of the photoresist. The EUV transmission at a 46.9 nm wavelength is reduced approximately by half every 10 nm thickness of PMMA [26]. It also should be taken into account that peak intensity is changing as a function of M sf . The advantage of obtaining high M sf with fractional Talbot lithography is opposed by the fact that intensity reduces with increasing M sf as shown in Fig. 5. The intensity values from simulations (blue squares) and developed depth of the photoresist from experiments (red circles) are both decreasing for higher M sf , as expected. The fractional patterning with spatial frequency multiplication will be useful for applications that require large area and high-resolution patterning, relaxing the fabrication constrains for the master mask. It is also convenient and efficient when various sizes have to be printed using one mask in EUV patterning. In conclusion, we have demonstrated a fractional Talbot lithography experimentally using highly coherent EUV illumination at a 46.9 nm wavelength generated with the capillary discharge Ne-like Ar laser. Spatial frequency multiplications up to 5x were achieved with pitches of 3 μm, 1.5 μm, 1 μm, and 600 nm, respectively, from the parent grating of 3 μm pitch. The grating mask is highly transmissive with a free standing structure fabricated by FIB milling. However, the line width roughness and the out-of-plane bending of the lines must be improved for better quality of the printing over large areas. We envision that a grating with a smaller pitch size could be used for further reduction of period and feature size, and the
6972
OPTICS LETTERS / Vol. 39, No. 24 / December 15, 2014
grating with narrower slits could be used to get higher M sf values, more than M sf 5x. In addition, the utilization of recently developed EUV lasers in the vicinity of 13 nm can decrease further the feature size achievable with this patterning approach [27]. This work is performed within the EU FP7 Erasmus Mundus Joint Doctorate Programme EXTATIC under framework partnership agreement FPA-2012-0033. The authors also would like to thank Elmar Neumann from the Helmholtz Nanoelectronic Facility at the Forschungszentrum Juelich for help with FIB milling. References 1. H. H. Solak, J. Phys. D 39, R171 (2006). 2. The International Technology Roadmap for Semiconductors, http://www.itrs.net (2013). 3. E. Ozbay, Science 311, 189 (2006). 4. T. F. Krauss, R. M. De La Rue, and S. Brand, Nature 383, 699 (1996). 5. M. Reed, J. Randall, R. Aggarwal, R. Matyi, T. Moore, and A. Wetsel, Phys. Rev. Lett. 60, 535 (1988). 6. P. Latimer and R. F. Crouse, Appl. Opt. 31, 80 (1992). 7. A. Isoyan, F. Jiang, Y. C. Cheng, F. Cerrina, P. Wachulak, L. Urbanski, J. Rocca, C. Menoni, and M. Marconi, J. Vac. Sci. Technol. B 27, 2931 (2009). 8. L. Wang, H. H. Solak, and Y. Ekinci, Nanotechnology 23, 305303 (2005). 9. H. H. Solak, C. Dais, and F. Clube, Opt. Express 19, 10686 (2011). 10. L. Urbanski, A. Isoyan, A. Stein, J. J. Rocca, C. S. Menoni, and M. C. Marconi, Opt. Lett. 37, 3633 (2012). 11. W. Li, V. M. Esquiroz, L. Urbanski, D. Patel, C. S. Menoni, M. C. Marconi, A. Stein, W. Chao, and E. H. Anderson, J. Vac. Sci. Technol. B 31, 06F604 (2013).
12. N. Guerineau, E. D. Mambro, and J. Primot, Opt. Express 11, 3310 (2003). 13. H. H. Solak and Y. Ekinci, J. Vac. Sci. Technol. B 23, 2705 (2005). 14. S. Danylyuk, P. Loosen, K. Bergmann, H.-S. Kim, and L. Juschkin, J. Micro-Nanolith. MEM. 12, 033002 (2013). 15. S. Brose, S. Danylyuk, L. Juschkin, C. Dittberner, K. Bergmann, J. Moers, G. Panaitov, St. Trellenkamp, P. Loosen, and D. Grützmacher, Thin Solid Films 520, 5080 (2012). 16. M. V. Berry and S. Klein, J. Mod. Opt. 43, 2139 (1996). 17. J. Wen, Y. Zhang, and M. Xiao, Adv. Opt. Photon. 5, 83 (2013). 18. J. T. Winthrop and C. R. Worthington, J. Opt. Soc. of Am. 55, 373 (1965). 19. V. Arrizón and J. Ojeda-Castañeda, Appl. Opt. 33, 5925 (1994) 20. D. D. Flanders, A. M. Hawryluk, and H. I. Smith, J. Vac. Sci. Technol. 16, 1949 (1979). 21. J. W. Goodman, Introduction to Fourier Optics, 2nd ed. (McGraw-Hill, 1996). 22. W. B. Case, M. Tomandl, S. Deachapunya, and M. Arndt, Opt. Express 17, 20966 (2009). 23. L. Urbanski, M. C. Marconi, L. M. Meng, M. Berrill, O. Guilbaud, A. Klisnick, and J. J. Rocca, Phys. Rev. A 85, 033837 (2012). 24. Y. Liu, M. Seminario, F. G. Tomasel, C. Chang, J. J. Rocca, and D. T. Attwood, Phys. Rev. A 63, 033802 (2001). 25. M. C. Marconi, J. L. A. Chilla, C. H. Moreno, B. R. Benware, and J. J. Rocca, Phys. Rev. Lett. 79, 2799 (1997). 26. B. L. Henke, E. M. Gullikson, and J. C. Davis, At. Data Nucl. Data Tables 54, 181 (1993). 27. D. Alessi, D. H. Martz, Y. Wang, M. Berrill, B. M. Luther, and J. J. Rocca, Opt. Lett. 35, 414 (2010).
