Investigating extreme ultraviolet radiation chemistry with first-principles quantum chemistry calculations

Abstract

In extreme ultraviolet (EUV) lithography, chemistry is driven by secondary electrons. A deeper understanding of these processes is needed. However, electron-driven processes are inherently difficult to experimentally characterize for EUV materials, impeding targeted material engineering. A computational framework is needed to provide information for rational material engineering and identification at a molecular level. We demonstrate that density functional theory calculations can fulfill this purpose. We first demonstrate that primary electron energy spectrum can be predicted accurately. Second, the dynamics of a photoacid generator upon excitation or electron attachment are studied with ab-initio molecular dynamics calculations. Third, we demonstrate that electron attachment affinity is a good predictor of reduction potential and dose to clear. The correlation between such calculations and experiments suggests that these methods can be applied to computationally screen and design molecular components of EUV material and speed up the development process.

High-Performance Chain Scissionable Resists for Extreme Ultraviolet Lithography: Discovery of the Photoacid Generator Structure and Mechanism

“…In spite of this high barrier, PAG 7 is still very active under EUV exposure. This clashes with the widespread perception that PAGs react almost exclusively with thermalized electrons to dissociate, and that higher energy processes can be safely discounted. ,, Clearly, this assumption fails here. To obtain a single parameter that estimates the overall rate of the dissociative electron transfer process (Scheme ), we combine the VEA and kinetic barrier of the N–O bond cleavage to create an empirical relationship with photospeed shown in Figure .…”

“…This clashes with the widespread perception that PAGs react almost exclusively with thermalized electrons to dissociate, and that higher energy processes can be safely discounted. 10,46,72 Clearly, this assumption fails here. To obtain a single parameter that estimates the overall rate of the dissociative electron transfer process (Scheme 7), we combine the VEA As can be seen, the correlation between this estimation of total activation energy and photospeed is excellent.…”

“…This trend has been verified experimentally ,, and computationally in e-beam lithography. Additionally, PAG reduction potentials have been shown to correlate linearly with their vertical electron affinities (VEAs) . Therefore, DFT calculations were first conducted to evaluate the VEAs of PAGs 3–8 .…”

“…While the acid generation efficiency of nonionic PAGs has been demonstrated to strongly correlate with the PAGs’ electron affinity under e-beam conditions, ,− such a trend has been consistently violated , when the exposure source has been changed to EUV. Many authors have recently speculated on the potential origin of this anomalous behavior, , but no substantial evidence has been provided. Because the initial step in the decomposition of nonionic PAGs is the formation of a radical anion, it has been exceptionally difficult to explore these multiconfigurational systems using traditional density functional theory (DFT).…”

Atomistic Modeling Approach for Predicting Association of Photoacid Generators in Extreme Ultraviolet Polymeric Photoresists

“…Chemical inhomogeneities in these mixtures are an aggravating factor that increases the likelihood of stochastic defects during lithography, and decreasing inhomogeneity is one way to decrease defectivity. As target critical dimensions continue to shrink, understanding the dispersibility of PAGs and PDQs in the polymer base is important to minimize defects that may arise from PAG agglomerates which may affect line edge roughness (LER) and stochastic defects. , Many experimental − and molecular modeling ,− studies have mostly concentrated on the processes occurring during and post-EUV exposure, e.g., to track chemical reactions that induce acid generation and estimate their effect on feature resolution and defects. Stochastic defects are inherent in the processing of photoresists even in the pre-EUV exposure conditions where the intermolecular interactions among the polymer base, PAGs, and PDQs can lead to nanosegregated domains whose sizes are in the same order of magnitude as that of the desired sub-10 nm features.…”