EUV Resists are Getting Exposed by More than EUV Light

Electrons from underneath the EUV resist are reducing dose-to-size by over 30%

Dose is one of the critical parameters for exposing resists in EUV lithography. A lower dose favors higher throughput, but aggravates stochastics due to low photon density [1]. However, a higher dose is not without complicating effects. First, resists have been shown to be thinned at higher EUV doses, due to outgassing [2]. Second, while the defect density is reduced by going to higher doses, the variability of defect density is increased [3]. The complicating factor for EUV is that it is not simply absorption that drives the resist chemical response, but the electrons released by the absorption [4]. Also, unlike the case for DUV, there is very little reflection of EUV light from the layers under the EUV resist, or by the EUV resist itself, since the refractive index is extremely close to 1 [5].

This year, papers from Tokyo Electron (TEL) [6] and Samsung/Synopsys [7] have revealed that EUV light that has not been absorbed by the resist but transmitted through to the underlayer can release electrons that can contribute significantly to the exposure of the resist. In particular, TEL found that dose to size for 28 nm pitch lines and spaces was reduced from 81.4 mJ/cm2 to 42.6 mJ/cm2, nearly 50%, by changing the underlayer. For 36 nm pitch hexagonal pillars, it was reduced from 88.3 mJ/cm2 to 59.6 mJ/cm2, or about a third. Samsung and Synopsys found that increasing underlayer thickness from 5 nm to 15 nm could reduce dose to size by about 15% for line pitches of 50 nm or larger. They attributed this to diffusion of “active particles” from the underlayer into the resist. Oddly, they describe these active particles to be molecules rather than secondary electrons. Secondary electrons are capable of diffusing over 10 nm [8]. A thicker underlayer provides more opportunity to release secondary electrons [9], whereas diffusion of an active chemical species should actually be less likely from chemical gradient considerations.

Diving deeper into the exposure from the underlayer, we can investigate the exposure of the bottom of the resist vs. the top. TEL found that the profile improved from being wider at the top to being less tapered, presumably due to more exposure of the lower portions of the resist [9]. We can also estimate the relative contribution of underlayer electron exposure vs. exposure from electrons resulting from resist absorption. Let us take the exposure from electrons from resist absorption in the reference case as A, and the exposure from electrons from the underlayer as B. When an underlayer change results in a reduction of dose to size, the exposure from resist absorption drops from A to rA (r<1). However, the total electron exposure is not reduced, but should be at least the same, A+B. Therefore, the exposure from underlayer electrons becomes A+B-rA, or (1-r)A+B. The ratio of underlayer exposure to resist absorption exposure goes from B/A to [(1-r)A+B]/(rA) = (1-r)/r + (1/r) B/A, which is a definite increase over the reference case. For example, for r=0.5, which is close to the 28 nm pitch case above, the ratio increases from B/A to 1+2 B/A. If B/A was originally 10% for the reference it becomes 120% with the underlayer change. This means the new underlayer’s electron exposure dominates the overall EUV resist exposure; the actual resist absorption actually becomes relatively minor. This could have significant effects on the resist profile, CD (critical dimension), and even the defectivity.

Since the underlayer electron exposure effectively originates from some distance below the resist layer, even if the image is focused within the resist layer, the underlayer electron exposure comes from a defocused image some distance below. This could be especially detrimental for High-NA EUV exposures, which are already limited by depth of focus. A larger relative contribution from the defocused underlayer electron exposure leads to reduced contrast in the total electron dose in the resist, which therefore becomes more sensitive to stochastic dose variations. A higher EUV dose also means a higher underlayer electron dose, which can now include a wider range of distances from underneath the resist as well as a wider range of values of local secondary electron yield. This can therefore be a new source of stochastic exposure variation in EUV lithography. The image blur is also effectively increased, due to the increased defocused and blurred exposure contribution from the underlayer’s scattered secondary electrons [10]. A larger blur increases stochastic defectivity since it becomes easier for a nanoscale region to erroneously cross the printing threshold [11]. Therefore, the underlayer’s electron yield and mean free path has become the most significant factor in EUV lithography performance.

References

[1] F. Chen, “Facing the Quantum Nature of EUV Lithography”.

[2] F. Chen, “EUV Resist Degradation with Outgassing at Higher Doses”.

[3] F. Chen, “Explaining ppm-level Stochastic Defectivity in a 3nm Via EUV Lithography Process”; F. Chen, “The Origin of Unstable EUV Process Yields”; F. Chen, “What Drives Stochastic Defectivity in EUV Lithography?”; H. Fukuda, Y. Momonoi, K. Sakai, “Estimating extremely low probability of stochastic defect in extreme ultraviolet lithography from critical dimension distribution measurement,” J. Micro/Nanolith. MEMS MOEMS 18, 024002 (2019).

[4] See, for example, I. Pollentier et al., “Unraveling the role of secondary electrons upon their interaction with photoresist during EUV exposure,” Proc. SPIE 10450, 104500H (2017).

[5] C. W. Maloney, B. W. Smith, “Longer Wavelength EUV Lithography (LW-EUVL)”.

[6] Y. Okumura et al., “Optimization of resist underlayer and novel development method for extreme ultraviolet,” Proc. SPIE 13983, 139832P (2026).

[7] S. Choi et al., “Understanding the Underlayer Effect using lithographic modeling,” Proc. SPIE 13983, 139831D (2026).

[8] A. Narasimhan et al., “What We Don't Know About EUV Exposure Mechanisms,” J. Photopolym. Sci. Tech. 30, 113 (2017).

[9] N. Miyahara et al., “Fundamentals of EUV stack for improving patterning performance,” Proc. SPIE 12498, 124981E (2023).

[10] F. Chen, “Explaining Metal Oxide Resist Bridging, Not Breaking, at Higher EUV Doses”.

[11] F. Chen, “How EUV Resist Blur and Dose-Induced Thinning Set the Stochastic Defectivity Floor”.