The study of photon shot noise effect in EUVL by using attenuated PSM

Abstract

Extreme ultraviolet lithography (EUVL), which uses 13.5 nm wavelength, is considered as the most promising technology for contact and hole (C/H) patterns of size below 27 nm. In particular, extreme ultraviolet (EUV) masks are considered important in facilitating mass production of EUVL technology. Unlike its predecessors, EUVL employs reflective mirrors, since EUV light is heavily absorbed by most materials. EUVL currently faces numerous challenges, including the mask shadowing effect and photon shot noise effect. The photon shot noise effect involves statistical fluctuations between the photon and photo resist (PR). This effect results in deterioration of the imaging performance. Photon shot noise is considered as a significant concern for EUVL owing to its energetic photons. The energy of EUV light is 14.3 times larger than that of ArF light, and thus, the number of photons required for reacting with the PRs is smaller in EUVL for the same dose, thus resulting in a stronger photon shot noise effect. To alleviate this effect, PR with higher quantum efficiency or a method to transfer more photons onto wafer are necessary. In this paper, we discuss the decrease in the photon shot noise effect by using EUVL phase shift mask (PSM). In chapter 2, we discuss photon shot noise and its mechanism. All light, including EUV, is a sum of photon energy. Photon shot noise originates from the discontinuous light flow energy that is similar to the shot gun form, and this effect has an inevitable light property, which leads to degraded stochastic imaging properties. EUV light source wavelength (13.5 nm) is 14 times smaller compared to ArF light source wavelength, and therefore, for the same amount of light energy, the EUV light source shows a much greater photon shot noise effect because its energy is 14 times higher than that of the ArF light source. In photolithography, this effect involves a reaction between the PR and particle, and so the weaker the statistical photon effect, the better the CD uniformity (CDU), line and contact edge roughness (LER and CER), and other stochastic imaging properties. Methods to decrease the photon shot noise effect include dose increase, new PR development, and rinse process development. But this research focuses on EUV masks. So that the effects of the mask on the photon shot noise effect are considered. In chapter 3, we explain stochastic simulation and the simulation parameters. There are two types of PR simulations, one is a probability-based model for numerical stochastic simulation and another is an analytical method for using determinism of continuum simulations. In this study, we used stochastic simulation since the photon shot noise effect in lithography is a probable effect between photons and PR. Also, the attenuated PSM used in this experiment is theoretically explained with respect to its properties, and the inserted PSM structure and properties are explained. The high numerical aperture (high-NA) in the exposure system of EUV lithography will be explained theoretically. Further, a high-NA system was applied in this study. In chapter 4, we present the proposed PSM used to simulate the mitigation of photon shot noise effect in the high-NA system. When 12 nm hp C/H patterns and 0.55 high-NA system is applied, the stochastic imaging properties such as CDU and CER were degraded by the photon shot noise effect. With 4x demagnification and 9° chief ray angle (CRA), compared to the 2% phase shift mask, which is the standard in this study, different phase shift masks of 6%, 12%, and 18% exhibited CDU mitigation of 6.1%, 15.9%, and 25.1% and CER mitigation of 21.6%, 31.9%, and 39.7%. With 8x demagnification and 6° CRA, compared to the standard 2% phase shift mask, the phase shift mask of 6%, 12%, and 18% showed a CDU mitigation of 4.1%, 10.0%, and 14.1%, and the CER mitigation of 16.5%, 28.3%, and 34.9%. Further, when setting the x-axis at 4x and y-axis at 8x demagnification, also known as anamorphic, the 6° CRA of a 2% phase shift mask compared to a 6%, 12%, and 18% phase shift mask showed a CDU mitigation of 8.1%, 14.2%, and 22.0%, and CER mitigation of 17.9%, 26.6%, and 28.8%. Through these results, the mitigation of photon shot noise effect will be proved from the reflectivity changes of the PSM, and the validity of the assumption discussed in chapter 2 will be assessed. In conclusion, the photon shot noise effect in EUVL can be effectively mitigated by controlling the reflectivity of the absorber stack of the attenuated PSM. The attenuated PSM can be an alternate technique for mitigating the photon shot noise effect.