Mask Materials and Designs for Extreme Ultra Violet Lithography
- 198 Downloads
Extreme ultra violet lithography (EUVL) is no longer a future technology but is going to be inserted into mass production of semiconductor devices of 7 nm technology node in 2018. EUVL is an extension of optical lithography using extremely short wavelength (13.5 nm). This short wavelength requires major modifications in the optical systems due to the very strong absorption of EUV light by materials. Refractive optics can no longer be used, and reflective optics is the only solution to transfer image from mask to wafer. This is why we need the multilayer (ML) mirror-based mask as well as an oblique incident angle of light. This paper discusses the principal theory on the EUV mask design and its component materials including ML reflector and EUV absorber. Mask shadowing effect (or mask 3D effect) is explained and its technical solutions like phase shift mask is reviewed. Even though not all the technical issues on EUV mask are handled in this review paper, you will be able to understand the principles determining the performance of EUV masks.
KeywordsExtreme ultraviolet lithography EUV mask Absorber materials Multilayer mirror Phase shift mask
2 EUV Mask Materials and Fabrication Process
3 Substrate Material
Mask MSFR (also referred to as slope error) is not a source of flare as it is in mirrors, but rather leads to intensity variation in the image plane and becomes a source of line edge roughness (LER) or image placement error . It was proposed that the frequency ranges for MSFR and HSFR should be 10−6/nm < f < 0.004/nm and 0.004/nm < f < 0.02/nm, respectively. The ML smoothing deposition technique reduces the mask HSFR, but it is not effective for MSFR. The SEMI standard specifies a HSFR of < 0.15 nm rms, and a local slope angle of the front surface < 1.0 mrad.
The front surface flatness of an as-mounted mask is influenced by the freestanding flatness of the substrate surfaces, and by how parallel they are to each other . The backside flatness of the mask also contributes since a flat-faced electrostatic chuck is the designated method of mounting a mask in an exposure tool . The SEMI P37 standard specifies the front (freestanding) and backside flatness values less than 30-nm peak to valley (P–V).
4 ML Mirror Layer
A capping layer is used for protecting the ML from oxidation since the reflectivity loss due to surface oxidation is expected for Si-terminated ML. Among the several candidate materials, Ru is extensively used as a capping layer for EUV masks owing to its suitability for EUV mask manufacturing [28, 29, 30, 31, 32]. Thin (~ 2 nm) Ru capping layer exhibits high etch selectivity to the absorber layer  and significant chemical stability under mask wet cleaning condition in addition to the minimum reflectivity loss when added on top of the ML. Moreover, the capping layer is essential for protecting the ML surface during pattern repair process using Focused Ion Beam (FIB) or E-beam .
5 Absorber Layer
According to the simulation, the use of TaN and Cr, which have a larger Δn (the difference between the real part of the refractive index and that of the vacuum), results in a larger shadowing effect . As the absorber thickness increases, the difference in CD for different materials decreases. Several researchers have investigated methods for reducing the shadowing effect by using thin absorber materials such as Ni, Pd, Pt, and Co, having high extinction coefficient for EUV wavelength [48, 55, 56, 61]. It should be noted that the difference caused by the shadowing effect can be compensated via mask biasing; however, it can ultimately limit the resolution [62, 63]. To obtain the highest image contrast, the absorber thickness must be sufficient so that the leakage light is as low as possible.
It is advisable to stop the absorber etch process with a sufficient selectivity to the capping layer. A high etch selectivity of 60:1 was reported against a Ru capping layer for TaN absorber etching with hexafluoric (SF6) chemistry . An etch bias of less than 5 nm was obtained that was independent of both structure and size, which contrasts with the results of the Cr absorber. Other important issues in the absorber etch process are the sidewall angle and line edge roughness (LER). Moreover, the absorber etch process preferably requires a minimal effect on the EUV reflectivity of the ML. Possible contributing factors for reflectivity loss are the surface oxidation or surface morphology change. Sufficient uniformity with only a small loss of reflectance has been reported with a Ta-based absorber etch process . Durability of ARC during etch is an important issue as well, because any reduction of the ARC thickness affects the DUV inspection contrast. Ru is a prominent capping material owing to its high etch selectivity against TaN absorber [32, 66].
As the ions of GFIS are generated in a field of a few atoms, the source size (Rs) of GFIS is significantly smaller than that of LMIS. As gas molecules are ionized in an electric field of particular intensity, the energy spread (ΔE) of GFIS is significantly smaller than that of LMIS. Therefore, the beam diameter (D) of GFIS is significantly smaller than that of LMIS .
6 Advanced Mask Structure for Better Imaging
Alternative mask structures can alleviate mask shadowing effect and eventually extend the resolution limit of EUV lithography. Various kinds of advanced EUV masks include (1) an etched ML binary mask, (2) an attenuated PSM (embedded or etched ML), (3) an alternating PSM (substrate or ML etched), and (4) a modified alternating PSM (double etched or with an absorber stack) [79, 80, 81, 82, 83, 84]. Most of the studies have been based on simulation, but some of the experimental results have revealed the possibility of practical implementation. One such achievement is the etched ML binary mask. Mo/Si ML (20 pairs instead of 40 pairs) was locally etched to form dark trenches, and its optical properties were investigated. H–V CD bias and best focus shift through pitch were improved compared to the standard mask, but some optical properties were degraded (e.g., higher mask error enhancement factor: MEEF). The most critical challenges of this structure are the manufacturability and cleanability . Based on the papers presented previously, the etched ML mask generally shows (1) larger depth of focus, (2) higher contrast, (3) no line width variation due to interference phenomenon, and (4) reduced H–V bias and image placement error (IPE) due to the shadow effect [86, 87].
One of the most effective resolution enhancement technologies in lithography is increasing the numerical aperture (NA) of the projection optics to collect more diffracted light from the mask. Therefore, it is necessary to adopt optical elements with an NA higher than 0.33 to extend the resolution below 8 nm in EUVL [93, 94]. A CRA of 6° and 4X demagnification cannot be maintained with an NA of 0.45 because the incident and reflecting light cones overlap. To separate the light cones, an increase of the CRA or demagnification factor is needed. With a demagnification factor greater than 4, the mask should be enlarged to prevent a reduced exposure field size, which occurs with a higher demagnification factor, resulting in throughput decrease [95, 96]. However, it is difficult to change the size of the mask because this requires a large change in the mask fabrication infrastructure. If the CRA is increased instead to separate two light cones, the mask 3D effect due to the absorber thickness and oblique incident light is intensified [97, 98]. Moreover, imbalance in 0th order diffractions, which is due to reflectivity variation according to the angle of incidence, results in a telecentricity error causing a pattern shift on the wafer [99, 100]. Since pattern shifts and mask 3D effects become more significant with a finer pattern pitch, modified designs of the absorber stack and ML mirrors are required to alleviate the above-mentioned problems .
The mask shadowing effect becomes even more significant due to the greater angle of incident light. This effect results in a smaller printed space on the wafer compared to the designed space on the mask. This in turn results in a horizontal–vertical critical dimension (H–V CD) bias and, eventually, a limitation in resolution. A high-NA EUV mask requires a thinner absorber, but this can result in the deterioration in the imaging properties due to the decreased mask contrast. A thin absorber material with a high extinction coefficient or phase shift absorber stack will help to alleviate the mask 3D effect with a high incident angle of the high-NA system.
EUVL is an extension of optical lithography that uses 13.5-nm-wavelength light. This wavelength requires major modifications in the standard optical lithography techniques, primarily due to the very strong absorption of EUV light by most materials. Refractive optics can no longer be used, necessitating the use of an ML mirror-based mask as well as an oblique incident angle of light. The imaging performance of the mask strongly depends on the mask structure and materials. The basic technologies related to the substrate, ML, absorber stack, mask patterning, cleaning, and metrology have recently shown significant improvements, but further systematic studies are needed to meet the requirements of 7 nm technology node applications. Resolution enhancement technology such as PSMs might be a solution to the technology’s extension, but this will require collaborative studies in relation to the simulation, metrology, and fabrication processes. In addition, many challenges still remain with regard to the mask infrastructure such as EUV mask fabrication, inspection, repair, pellicle integration, and handling of pelliclized masks .
I would like to thank all the authors of the technical papers referenced in this review paper. The author is indebted to all the students and colleagues for their dedicated assistance. This research was supported by the Commercialization Promotion Agency for R&D Outcomes (COMPA), funded by the Ministry of Science and ICT (MSIT) (Grant No. 2017K000389).
- 1.Chen, A., Miyazaki, J.: EUV lithography insertion for high volume manufacturing: status and outlook. In: Proceedings of IEEE Electron Devices Technology and Manufacturing Conference, p. 77. IEEE Electron Devices Society, Toyama, Japan (2017)Google Scholar
- 4.van Es, R., van de Kerkhof, M., Jasper, H., Levasier, L., Peeters, R.: EUV lithography industrialization progress. Proc. SPIE 10450, 1045003 (2017)Google Scholar
- 5.Fomenkov, I., Brandt, D., Ershov, A., Schafgans, A., Tao, Y., Vaschenko, G., Rokitski, S., Kats, M., Vargas, M., Purvis, M., Rafac, R., La Fontaine, B., De Dea, S., LaForge, A., Stewart, J., Chang, S., Graham, M., Riggs, D., Taylor, T., Abraham, M., Brown, D.: Light sources for high-volume manufacturing EUV lithography: technology, performance, and power scaling. Adv. Opt. Technol. 6, 173 (2017)Google Scholar
- 6.Yabu, T., Kawasuji, Y., Hori, T., Okamoto, T., Tanaka, H., Miyao, K., Ishii, T., Watanabe, Y., Yanagida, T., Shiraishi, Y., Abe, T., Kodama, T., Nakarai, H., Yamazaki, T., Itou, N., Saito, T., Mizoguchi, H.: Key components development progress updates of the 250W high power LPP-EUV light source. Proc. SPIE 10450, 104501C (2017)Google Scholar
- 7.Yan, P.-Y.: Handbook of Photomask Manufacturing Technology, p. 234. CRC Press, Boca Raton (2005)Google Scholar
- 8.Hector, S.: Standards for EUV Masks. SEMI EUV Mask Standards Meeting. SEMATECH, San Francisco, California (2005)Google Scholar
- 11.Hector, S.: Standards for EUV Masks. EUV Mask Workshop. SEMATECH, Miyazaki, Japan (2004)Google Scholar
- 12.Gullikson, E., Blaedel, K., Larson, C., Baker, S.L., Taylor, J.S.: EUV scattering from mask substrate roughness. 1st EUVL Symposium. SEMATECH, Dallas, Texas (2002)Google Scholar
- 14.Turley, C., Rankin, J., Cehn, X., Ballman, K., Lee, C.A., Dunn, T.: EUV mask flatness compensation strategies and requirements for reticle flatness, scanner optimization and E-beam writer. Proc. SPIE 10450, 104500A (2017)Google Scholar
- 16.Chen, X., Turley, C., Rankin, J., Brunner, T., Gabor, A.: Minimizing wafer overlay errors due to EUV mask non-flatness and thickness variations for N7 production. Proc. SPIE 10143, 101431F (2017)Google Scholar
- 19.Yan, P.-Y.: Handbook of Photomask Manufacturing Technology, p. 238. CRC Press, Boca Raton (2005)Google Scholar
- 20.Melvin, L.S., Kandel, Y., Isoyan, A., Gao, W.: Individual multilayer reflectance and near field image formation in an EUV reticle. Proc. SPIE 10450, 104500F (2017)Google Scholar
- 21.Onoue, T., Shoki, T., Horikawa, J.: Progress of EUV blanks development. EUVL Symposium, S2.1. EIDEC, Hiroshima, Japan (2016)Google Scholar
- 22.Tomofuji, T., Kandaka, N., Komiya, T., Shiraishi, M., Murakami, K.: Mo/Si multilayer(ML) mirror depositied with ion beam sputtering using Kr gas. 3rd International EUVL Symposium, p. 937. SEMATECH, Miyazaki, Japan (2004)Google Scholar
- 31.Dattilo, D., Dietze, U., Hsu, J.-W.: Ruthenium capping layer preservation for 100X clean through pH driven effects. Proc. SPIE 9635, 96351B (2015)Google Scholar
- 38.Mirkarimi, P.B., Spiller, E.A., Baker, S.L., Sperry, V.L., Stearns, D.G., Gullikson, E.M.: Developing a viable multilayer coating process for extreme ultraviolet lithography reticles. J. Microlithogr. Microfabr. Microsyst. 3(1), 139–145 (2004)Google Scholar
- 40.Kearney, P., Ma, A., Jeon, C.U., Uno, T., Beier, B.: Defect mitigation and reduction in EUVL mask blanks. 5th EUVL Symposium, p. 734. SEMATECH, Barcelona, Spain (2006)Google Scholar
- 42.Rastegar, A., Eichenlaub, S., Popp, H., Goncher, K., Marmillion, P.: Removing sub-50nm particles during blank substrate cleaning. Solid State Technol. 49, 47 (2006)Google Scholar
- 44.Mirkarimi, P.B., Spiller, E., Baker, S.L., Stearns, D.G., Robinson, J.C., Olynick, D.L., Salmassi, F., Liddle, J.A., Liang, T., Stivers, A.R.: A silicon-based, sequential coat-and-etch process to fabricate nearly perfect substrate surfaces. J. Nanosci. Nanotechnol. 6, 28 (2006)Google Scholar
- 48.Philipsen, V., Luong, K.V., Hendrickx, E., Erdmann, A., Xu, D., Evanschitzky, P., Kruijs, R.W., Edrisi, A., Scholze, F., Laubis, C., Irmscher, M., Naasz, S., Reuter, C.: Mitigating EUV mask 3D effects by alternative metal absorbers. EUVL Symposium, S4.2. EIDEC, Hiroshima, Japan (2016)Google Scholar
- 49.Philipsen, V., Luong, K.V., Souriau, L., Sanchez, E.A., Adelmann, C., Laubis, C., Scholtze, F., Kruemberg, J., Reuter, C., Hendrickx, E.: Single element and metal alloy novel EUV mask absorbers for improved imaging. Proc. SPIE 10450, 104500G (2017)Google Scholar
- 53.Green, M., Choi, Y., Ham, Y., Kamberian, H., Progler, C., Tseng, S.-E., Chiou, T.-B., Miyazaki, J., Lammers, A., Chen, A.: EUV mask manufacturing readiness in the merchant mask industry. Proc. SPIE 10450, 1045005 (2017)Google Scholar
- 55.Wood, E.O., Raghunathan, S., Mangat, P., Philipsen, V., Luong, V., Kearney, P., Verduijn, E., Ku-mar, A., Patil, S., Laubis, C., Soltwisch, V., Scholze, F.: Alternative materials for high numerical aperture extreme ultraviolet lithography mask stacks. Proc. SPIE 9422, 94220I (2015)Google Scholar
- 57.Philipsen, V., Luong, K.V., Souriau, L., Hendrickx, E., Erdmann, A., Xu, D., Evanschitzky, P., van de Kruijs, R.W.E., Edrisi, A., Scholze, F., Laubis, C., Irmscher, M., Naasz, S., Reuter, C.: Reducing EUV mask 3D effects by alternative metal absorbers. Proc. SPIE 10143, 1014310 (2017)CrossRefGoogle Scholar
- 64.Mangat, P.J.S., Hector, S.D., Thompson, M.A., Dauksher, W.J., Cobb, J., Cummings, K.D., Mancini, D.P., Resnick, D.J., Cardinale, G., Henderson, C., Kearney, P., Wedowski, M.: Extreme ultraviolet lithography mask patterning and printability studies with a Ta-based absorber. J. Vac. Sci. Technol. B 17(6), 3029 (1999)CrossRefGoogle Scholar
- 70.Lee, S.-Y., Kim, G.-B., Sim, H.-S., Lee, S.-H., Kim, H.-S., Lee, J.-H., Seo, H.-S., Han, H.-S., Kim, S.-S., Moon, S.-Y., Woo, S.-G., Bozak, R., Dinsdale, A., Robinson, T., Lee, D., Cho, H.K.: Analysis of process margin in EUV mask repair with nano-machining. Proc. SPIE 7122, 71222I (2008)CrossRefGoogle Scholar
- 82.Erdmann, A., Xu, D., Evanschitzky, P., Luong, V., Philipsen, V., Hendrickx, E.: Characterization and mitigation of 3D mask effects in EUV lithography. EUVL Symposium, S2.1. EIDEC, Hiroshima, Japan (2016)Google Scholar
- 87.Kamo, T., Takai, K., Iida, N., Morikawa, Y., Hayashi, N., Watanabe, H.: Evaluation of etched multilayer mask for 0.33NA EUVL extension. EUVL Symposium, S4.4. EIDEC, Hiroshima, Japan (2016)Google Scholar
- 88.Yan, P.-Y.: Handbook of Photomask Manufacturing Technology, p. 265. CRC Press, Boca Raton (2005)Google Scholar
- 91.Lee, J.U., Jeong, S.J., Hong, S., Lee, S.M., Ahn, J.: Imaging performance of attenuated phase-shift mask using coherent scattering microscope. Proc. SPIE 9048, 90481X (2014)Google Scholar
- 94.van Schoot, J., Troost, K., Bornebroek, F., van Ballegoij, R., Lok, S., Krabbendam, P., Stoeldraijer, J., Loopstra, E., Benschop, J., Finders, J., Meiling, H., van Setten, E., Kneer, B., Kuerz, P., Kaiser, W., Heil, T., Migura, S., Neumann, J.T.: High-NA EUV lithography enabling Moore’s law in the next decade. Proc. SPIE 10450, 104500U (2017)Google Scholar
- 98.Pirati, A., van Schoot, J., Troost, K., van Ballegoij, R., Krabbendam, P., Stoeldraijer, J., Loopstra, E., Benschop, J., Finders, J., Meiling, H., van Setten, E., Mika, N., Driedonkx, J., Stamm, U.: The future of EUV lithography: enabling Moore’s law in the next decade. Proc. SPIE 10143, 101430G (2017)CrossRefGoogle Scholar
- 100.Hosler, E.R., Thiruvengadam, S., Cantone, J.R., Civay, D.E., Schroeder, U.P.: EUV and optical lithographic pattern shift at the 5nm node. Proc. SPIE 9776, 977616 (2015)Google Scholar
- 101.Wood, O., Wong, K., Parks, V., Kearney, P., Ilse, J.M., Luong, V., Philipsen, V., Faheem, M., Liang, Y., Kumar, A., Chen, E., Bennett, C., Fu, B., Gribelyuk, M., Zhao, W., Mangat, P., der Heide, P.V.: Improved Ru/Si multilayer reflective coatings for advanced extreme ultraviolet lithography photomasks. Proc. SPIE 9776, 977619 (2016)CrossRefGoogle Scholar