- Email: [email protected]
Impact of transmission non-uniformity of a wrinkled EUV pellicle for N5 patterning under various illuminations
Accepted Manuscript Impact of transmission non-uniformity of a wrinkled EUV pellicle for N5 patterning under various illuminations
In-Seon Kim, Guk-Jin Kim, Michael Yeung, Eytan Barouch, HyeKeun Oh PII: DOI: Reference:
S0167-9317(17)30040-0 doi: 10.1016/j.mee.2017.01.031 MEE 10458
To appear in:
Microelectronic Engineering
Received date: Revised date: Accepted date:
14 October 2016 19 January 2017 25 January 2017
Please cite this article as: In-Seon Kim, Guk-Jin Kim, Michael Yeung, Eytan Barouch, Hye-Keun Oh , Impact of transmission non-uniformity of a wrinkled EUV pellicle for N5 patterning under various illuminations. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Mee(2017), doi: 10.1016/j.mee.2017.01.031
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Impact of transmission non-uniformity of a wrinkled EUV pellicle for N5 patterning under various illuminations a
In-Seon Kim, a Guk-Jin Kim, b Michael Yeung, c Eytan Barouch, and a Hye-Keun Oh† Applied Physics Department, Hanyang University, Ansan, Korea. b Fastlitho, SanJose, CA, USA. c Boston University, Boston, MA, USA. Email: †[email protected]
RI
PT
a
Abstract
SC
To solve the defect problem during the exposure process of EUV lithography (EUVL), an EUV pellicle is suggested as a solution. Even though use of an EUV pellicle is considered an
NU
essential solution for defect control during the exposure process, it is not ready for application in real commercial processes due to how difficult it is to manufacture. Due to
MA
tight requirements for an EUV pellicle, flawless fabrication is impossible and deformations such as a wrinkle would result in serious patterning problems. These deformations lead to critical dimension (CD) non-uniformity due to the non-uniform transmission distribution
D
caused by a pellicle wrinkle. In this paper, we discuss the impact of transmission non-
PT E
uniformity caused by a wrinkled pellicle. When we treat the effects of a wrinkled pellicle, we considered off-axis-illumination, which is a promising resolution enhancement technology. By shrinking the target pattern size down to 7 nm nodes (N7) or 5 nm nodes (N5), a flexible
CE
illumination system is required to enhance the resolution ability. With dipole illuminations, resolution is varied by partial coherence and by not only the beam size at the pellicle, but also
AC
by the incident angle distribution. For these reasons, a non-uniform intensity distribution caused by a wrinkled pellicle is modified with changes in off-axis illumination conditions, including spatial coherence and the aperture shape. For N5 patterning, the CD non-uniformity of 0.2 nm occurs at 1.7% transmission variation (2-pass). However, the intensity distribution is varied under various illumination conditions in spite of the same pellicle conditions. The intensity non-uniformity goes up with increasing spatial coherence. Even though feature conditions of the wrinkled pellicle are the same, the transmission non-uniformity is changed with illumination conditions. Thus, the allowable feature conditions of the wrinkled pellicle may be changed. For achieving good CD uniformity, the allowable limit of the pellicle
ACCEPTED MANUSCRIPT
wrinkle is carefully studied because the impact of the wrinkled pellicle varies with the illumination conditions.
AC
CE
PT E
D
MA
NU
SC
RI
PT
Keywords: EUV lithography; EUV pellicle; pellicle; pellicle deformation; Transmission non-uniformity; CD uniformity
ACCEPTED MANUSCRIPT
1. INTRODUCTION EUV lithography is one of the more promising technologies used to fabricate integrated chips with 1X nm dimensions and below [1, 2]. EUV lithography has an exceptional capability to produce sophisticated logic patterns because of the high resolution of UV light. However, some obstacles still need to be solved to allow for high volume manufacturing
[3]
. Among
various problems, defect control is a serious problem that must be solved before mass
pellicle is suggested as a solution
[2-4]
PT
production can take place. To solve the defect problem during the exposure process, an EUV . The EUV pellicle is a thin membrane that protects the
RI
mask from contamination by wrapping the mask. Even though the EUV pellicle is considered
SC
as an essential solution for defect control during exposure process, it is not ready for application in a real-world process because it is hard to precisely manufacture. Due to tight
NU
requirements for the EUV pellicle [5], flawless fabrication is impossible and it is very fragile. Deformations, such as a wrinkle, would cause a non-uniform transmission distribution involving the critical dimension (CD) [5-7]. CD non-uniformity is a critical problem for device
MA
production, thus the transmission non-uniformity study of a wrinkled pellicle is very important to realizing practical applications of the pellicle. By shrinking the target pattern size down to 5 nm nodes (N5), a conventional illumination [2, 8-10]
. In order to extend the resolution, we need a high NA system because it is
PT E
criterion
D
system with a 0.33 numerical Aperture (NA) meets the resolution limit based on Rayleigh
impossible to develop N5 pattering with a 0.33 NA system owing to the resolution limitation. For N5 patterning, a high 0.55 NA system was suggested, with a large magnification in order
CE
to avoid a serious shadow effect
[8-13]
. In spite of the high NA system, off-axis illumination
(OAI) technology is also required to enhance the resolution ability
[8-11]
. For large patterns,
AC
circular shaped illumination is usually used, but dipole illumination with small spatial coherence can achieve high contrast with fine patterns. Off-axis illumination technology can enhance the image quality of patterning but it can also affect wrinkled pellicle imaging. With dipole illumination, two poles of illumination are incident on the mask separately and they have a different incident angle distribution. Due to the incident angle distribution, the illumination source size at the pellicle level is varied. Therefore, the non-uniform intensity distribution caused by a wrinkled pellicle changes with variation in off-axis illumination conditions such as the aperture shape and spatial coherence. In this paper, we discuss the wrinkled pellicle effect for N5 pattering. In order to understand the influence of the wrinkled
ACCEPTED MANUSCRIPT
pellicle for the N5 pattering environment, we will treat not only the various illumination conditions but also various pattern sizes.
2. OPTICAL MODELING FOR WRINKLED PELLICLE WITH OAI As mentioned in the introduction section, a pellicle is easily deformed due to its weak structure and such deformation can lead to transmission non-uniformity. In order to calculate the image distortion caused by a wrinkled pellicle, we will optically model the wrinkled
PT
pellicle for calculation of transmission variation. First, we express the wrinkled depth of pellicle as a sinusoidal function, as shown in the schematic diagram of optical modeling in
PT E
D
MA
NU
SC
RI
Fig. 1 [14].
CE
Figure 1. Schematic diagram of the optical modeling for a wrinkled pellicle.
Height of wrinkled pellicle:
Eq. 1
AC
Since we can express the wrinkle height as a cosine function, we can calculate other feature conditions of the wrinkled pellicle. By differentiating the height function, we obtain a local slope at each position that represents the tangent value of the local tilt angle. Therefore, the arc tangent value of the local slope is a local incident angle, as shown in Eq. 2. Local tilt angle of wrinkled pellicle:
Eq. 2
With Eqs. 1 and 2, we can express the local tilt angle distribution as functions of the period and amplitude of the wrinkle. To calculate the optical path at the pellicle, we should know the incident angle for each position. The actual incident angle at the pellicle level is the sum of the local tilt angle of wrinkle and the angle distribution caused by OAI. If we use OAI technology, the incident angle is varied with aperture conditions. As shown in Fig. 2, EUV is
ACCEPTED MANUSCRIPT
obliquely incident at a 6 degree angle due to reflective optics and is focused at the mask and wafer
[2, 10]
. The incident angle distribution at mask level varies with spatial coherence, as
shown in Fig. 2(b). With 6 degrees of oblique incidence, the incident angle distribution caused by OAI can be expressed as Eq. 3. When we express the incident angle distribution, .
MA
NU
SC
RI
PT
we consider dipole illumination with spatial coherences (
AC
CE
PT E
D
(a)
(b) Figure 2. Light path for EUV lithography:
(a) EUV light path at projection optics and (b) illumination cone with dipole aperture at mask level.
Incident angle of Pole A: Incident angle of Pole B: The actual incident angle (
Eq. 3
at the wrinkled pellicle is a sum of Eqs. 2 and 3. With the
actual incident angle, the optical path at the pellicle (
can be calculated, as shown in Fig.
1 and Eq. 4. The absorption is exponentially proportional to the optical path and one pass transmission is presented in Eq. 4.
ACCEPTED MANUSCRIPT
one pass transmission of the wrinkled pellicle: Eq. 4 In Eq. 4, k is the extinction coefficient of the pellicle, dT is the optical path at the pellicle (shown in Fig. 1), λ is the wavelength of light, do is the thickness of the pellicle and θi is the actual incident angle. With this equation, we calculate the intensity distribution after passing
PT
through the pellicle once. As represented in Fig. 2(a), EUV light passes through the pellicle twice, and the beam position at the pellicle level is shifted due to the reflective optics and
RI
oblique incidence. Therefore, we should consider the shift caused by oblique incidence when we calculate the intensity distribution after passing through the pellicle twice. With
pellicle, which will be discussed in the next section.
SC
consideration of oblique incidence, we calculate the image distortion caused by the wrinkled
NU
3. CHARACTERISTICS OF TRANSMISSION NON-UNIFORMITY In this section, we will discuss image distortion caused by the wrinkled pellicle under
MA
various illumination conditions. When we calculate the transmission non-uniformity with the wrinkled pellicle, we consider the ideal case of the EUV pellicle. The ideal EUV pellicle was suggested by Intel first
[15]
. The ideal pellicle has a very simple structure,
D
which can achieve 90% transmission. The specifications of the ideal EUV pellicle are
PT E
summarized in Table 1. During the past few years, many different pellicle structures have been presented, but research regarding an optimized structure is still ongoing. Therefore, we use the ideal case at this time.
CE
Table 1. Simulation conditions of the EUV pellicle
c-Si 50 0.9991+0.0018i
AC
Material Thickness (nm) Refractive index (n+ik)[16]
3.1. Periodicity of transmission non-uniformity caused by oblique reflection. When the EUV light passes through the wrinkled pellicle, the incident angle varies with position. Due to the variation of incident angle, the optical path is also varied. By increasing the incident angle, the optical path is increased and thus causes a larger transmission loss, as shown in Fig. 3. As a result, the absorption is low at the top and bottom of a wrinkle, while the absorption at the sides of wrinkle is relatively higher due to a larger tilt angle and longer optical path.
PT
ACCEPTED MANUSCRIPT
RI
Figure 3. Incident angle and transmission variation of wrinkled pellicle.
SC
Owing to the transmission variation of a wrinkled pellicle, a wrinkled pellicle causes a non-uniform intensity distribution. One-pass transmission variation becomes very serious with a large tilt angle for the wrinkle. However, 2-pass transmission variation can be varied
NU
with period in spite of the same maximum local tilt angle because EUV light passes through the pellicle twice and incident light and reflected light pass at different positions due to
MA
oblique reflection. By considering transmission variation and passing position shift, we can
AC
CE
PT E
D
predict the cases of largest and smallest transmission difference, as shown in Fig. 4.
(a)
RI
PT
ACCEPTED MANUSCRIPT
PT E
D
MA
NU
SC
(b)
(c)
Figure 4. Periodicity of transmission non-uniformity of a wrinkled pellicle: (a) Conditions of the largest and smallest transmission difference cases, (b) intensity distribution
CE
after passing the pellicle twice, and (c) transmission non-uniformity as a function of the period of the wrinkle with the same maximum local tilt angle.
AC
In Fig. 4, graphs (a) and (b) show two cases of transmission variation. EUV lithography uses reflective optics with oblique incidence. Reflective optics and oblique incidence cause beam shift, as mentioned in the previous section. When the EUV light goes from a dark region to another dark region, the 2-pass transmission is a minimum. On the contrary, if the EUV light goes from a bright region to a bright region, the transmission is a maximum. On the left hand side of Fig. 4(a), the EUV light goes from a dark region to another dark region and a bright region to another bright region. Thus, the transmission difference will be large for this case. In contrast, when the EUV light goes from the bright region to dark region as in the case on the right hand side, the transmission difference is smaller than that of the left hand side case. As a result, the intensity distribution after passing through the pellicle twice
ACCEPTED MANUSCRIPT
is different for the two cases, as shown in Fig. 4(b). This indicates that the transmission nonuniformity is different with a different period of the wrinkle, despite having the same maximum local tilt angle. Figure 4(c) shows transmission non-uniformity as a function of the wrinkle period with a 300 mrad maximum local tilt angle. In Fig. 4(c), it is evident that transmission nonuniformity fluctuates with the period of the wrinkle. When the beam shift due to oblique
PT
reflection is a multiple of a half-period of the wrinkle, the transmitted light non-uniformity is at its largest. However, when the shift is a multiple of a half-period plus or minus a
RI
quarter period due to a different period, the 2-pass intensity non-uniformity is at its smallest.
SC
With consideration of the shift due to oblique incidence, the last case of periodicity is a wrinkle with a 1.05 mm period. Therefore, if the period is larger than about 1 mm, the periodicity of transmission non-uniformity does not appear.
NU
3.2. Beam size effect of partial coherent light.
Until now, we have explored pellicle imaging variation caused by reflective optics. Now,
MA
we will discuss the wrinkled pellicle imaging variation with partial coherent light. For resolution enhancement, various illumination aperture shapes and spatial coherences are used. By increasing the partial coherence, the beam size of the illumination cone at
D
pellicle will be increased. Owing to the variation of beam size, wrinkled pellicle imaging
PT E
can be varied. With dipole illumination, the intensity distribution after passing the pellicle
AC
CE
twice is shown in Fig. 5.
Figure 5. Wrinkled pellicle imaging with various illumination conditions.
ACCEPTED MANUSCRIPT
In Fig. 5, results show the intensity distribution after passing through the pellicle twice with various dipole sources. The illumination aperture shape is a dipole and the outer sigma is 0.9. As shown in Fig. 5, the intensity distribution caused by a wrinkled pellicle is varied with spatial coherence and the large transmission non-uniformity with a small dipole source (σin=0.8) is significantly decreased with a large dipole source (σin=0.2). In order to understand the beam size effect of a wrinkled pellicle, we separated the range of
PT
the wrinkle period. Figure 6 shows a 0.55 NA system at mask level. The suggested 0.55 NA system has an 8X magnification and the stand-off height between the mask and pellicle is
RI
fixed as 2.5 mm [5, 10]. With consideration of this structure, the maximum beam size with the
SC
0.55 NA system is about 350 μm. If the period of the wrinkle is larger than 1 mm, the distance between the dark region and bright region is far compared with the maximum beam size. As a result, the transmission non-uniformity caused by a wrinkled pellicle is decreased
PT E
D
MA
NU
in the large wrinkle region, as shown in Fig. 7.
AC
CE
Figure 6. Illumination system at the mask level with a 0.55 NA system.
ACCEPTED MANUSCRIPT
(b)
SC
RI
PT
(a)
Figure 7. Beam size effect of OAI with a wrinkled pellicle:
NU
(a) Transmission non-uniformity variation with various periods of wrinkle. (b) Transmission non-uniformity difference between a small and large wrinkle.
MA
Figure 6 shows the beam size effect of OAI with a wrinkled pellicle. If the period of the wrinkle is small, the transmission non-uniformity varies with aperture conditions due to the beam size effect. However, the difference of transmission non-uniformity with illumination
D
conditions is decreased as the period of the wrinkle is increased. Thus, if the period of the
PT E
wrinkle is much larger than the maximum beam size, the transmission non-uniformity is independent of aperture conditions. As can be seen in Fig. 6(b), the transmission nonuniformity as a function of the maximum local tilt angle also varies with illumination
CE
conditions in the small wrinkle region, but it is independent of illumination conditions in the large wrinkle region. With this criterion, the CD non-uniformity caused by the wrinkled
AC
pellicle may differ with illumination conditions in the small verses large wrinkle regions. This will be discussed in the next section.
4. CD NON-UNIFORMITY CAUSED BY A WRINKLE PELLICLE WITH N5 PATTERNS. After the EUV light passes through the pellicle twice, the wrinkled pellicle leads to a nonuniform intensity distribution, which causes CD non-uniformity. The CD difference between the bright region and dark region is linearly proportional to the intensity difference [14]
. In this section, we will measure the CD non-uniformity as a function of transmission
non-uniformity and then discuss the CD non-uniformity of N5 patterns caused by a wrinkled pellicle.
ACCEPTED MANUSCRIPT
In order to know the impact of the wrinkled pellicle for N5 patterns, we consider an isoline with a dense period pattern and N5 specifications, and measure CD variation as a function of transmission non-uniformity with 2 cases of dipole illumination conditions. Table 2. Simulation conditions for the CD non-uniformity characteristics study [16, 17].
EUVL suite (Fastlitho Inc.)
Simulation tool Iso-line
5
(nm)
Dense line & space
11 − 15
Tantalum boron nitride (TaBN)
Thickness
70 nm
Refractive index (n+ik) Small dipole
conditions
Large dipole
NU
Illumination
0.9165 + 0.0438i
SC
conditions
Material
RI
Absorber
PT
Target CD
The simulation conditions are summarized in Table. 2. With these conditions, we measured
MA
CD variation with a non-uniform intensity distribution and CD non-uniformity as a function of 2-pass transmission non-uniformity, as depicted in Fig. 8. In Fig. 8, graph (a) is the result
AC
CE
PT E
D
with the large dipole and graph (b) is the result with the small dipole.
(a)
RI
PT
ACCEPTED MANUSCRIPT
(b)
SC
Figure 8. CD non-uniformity as a function of transmission non-uniformity for N5 patterns.
If we define the slope value as a sensitivity, sensitivities are different not only with respect
CD uniformity budget is under 0.2 nm
[5]
NU
to pattern size, but also with respect to illumination conditions. For the EUV pellicle, the . To obtain 0.2 nm CD uniformity, the 2-pass
transmission non-uniformity should be smaller than 1.7% for all patterns, regardless of
MA
illumination conditions. The allowable transmission non-uniformity for N5 patterns is the same as for two dipole illuminations, but the intensity distribution caused by a wrinkled
D
pellicle varies with illumination conditions, as mentioned in the previous section. The CD non-uniformity as a function of maximum local tilt angle of the wrinkle is summarized in
AC
CE
PT E
Fig. 9.
Figure 9. CD non-uniformity as a function of the maximum local tilt angle of the wrinkle with dipole illumination.
In Fig. 9, gray lines are the results of the large dipole and black lines are that of the small dipole. The CD non-uniformity results in the small wrinkle region are expressed with solid lines and results in the large wrinkle region are described with dashed lines. As mentioned
ACCEPTED MANUSCRIPT
in Section 3, the transmission non-uniformity decreases with large spatial coherence in the small wrinkle region. As a result, the CD non-uniformity of the large dipole is smaller than that of the small dipole in the small wrinkle region. To obtain 0.2 nm CD uniformity, the maximum local tilt angle of the wrinkle should be smaller than 420 mrad with small dipole illumination for all N5 patterns. In the case of large dipole illumination, the CD uniformity variation is smaller than that of the small dipole and the requirement of maximum local tilt
PT
angle is relaxed to about 520 mrad. Thus, a large dipole gives a large allowable range of wrinkle compared to small dipole illuminations. In contrast to the small wrinkle, the
RI
transmission non-uniformity caused by a wrinkled pellicle is independent of the illumination
SC
conditions. Due to similar transmission variation and the same requirements of transmission non-uniformity, the requirement of maximum local tilt angle is almost the same for two dipole illuminations in the large wrinkle region, and the shared requirement of maximum
NU
local tilt angle is about 600 mrad.
5. CONCLUSION
MA
The EUV pellicle is an essential solution for the protection of a mask and it should have an extremely thin thickness for high transmission. Due to the weak structure of the pellicle, a wrinkled pellicle is an unavoidable problem. A pellicle wrinkle leads to a non-uniform
D
intensity distribution caused by transmission variations. Owing to the reflective optics and
PT E
oblique incidence of EUVL, the intensity non-uniformity varies according to the period of the wrinkle after passing the pellicle twice. With the consideration of oblique reflection and transmission distribution of the wrinkle, the last case of a large transmission difference is
CE
the wrinkle with a period of 1 mm. If the period of the wrinkle is larger than about 1 mm, the worst case of transmission non-uniformity is avoided completely. In the large wrinkle
AC
region, local intensity variation is relatively independent of not only the reflective optics, but also the illumination conditions of OAI. As the beam size varies at the pellicle, the intensity non-uniformity also varies, and it is increased with highly coherent light and a small wrinkle. However, the intensity non-uniformity is independent of illumination conditions in the large wrinkle region. With dipole illumination, the CD non-uniformity of 0.2 nm occurs with a transmission variation (2-pass) of 1.7%, regardless of illumination conditions. However, the allowable feature condition of a wrinkled pellicle is different due to different pellicle imaging results. For N5 patterning, the requirements of a maximum local tilt angle are 420 mrad for a small dipole and 520 mrad for a large dipole illumination.
ACCEPTED MANUSCRIPT
ACKNOWLEDGEMENT This research was supported partly by Samsung electronics mask development team and partly by the Nano Material Technology Development Program through the National Research Foundation of Korea (NRF), and is funded by the Ministry of Education, Science
AC
CE
PT E
D
MA
NU
SC
RI
PT
and Technology (No. 2015M3A7B7045353).
ACCEPTED MANUSCRIPT
REFERENCES 1. L. Wilson, “International Technology Roadmap for Semiconductors (ITRS)”, Semiconductor Industry Association (2013). 2. Vivek Bakshi, EUV Lithography, Bellingham, Washington, SPIE Press (2009). 3. Kurt Ronse, “Closing address”, International symposium on Extreme Ultraviolet lithography, Maastricht, Netherlands (2015).
PT
4. Harry J. Levinson, Principle of lithography, Bellingham, Washington, SPIE press (2005). 5. Derk Brouns, "Pellicle HVM specifications". Pellicle Technical Meeting Group, San Jose,
RI
USA. 2016.
SC
6. Luigi Scaccabarozzi, Dan Smith, Pedro Rizo Diago, Eric Casimiri, Nina Dziomkina, and Henk Meijer, “Investigation of EUV pellicle feasibility”, Proc. of SPIE, 8679, 867904 (2013).
NU
7. Dario L. Goldfarb, Max O. Bloomfield, and Matthew Colbum, “Thermo-mechanical behavior of EUV pellicle under dynamic exposure conditions”, Proc. of SPIE, 9776, 977621
MA
(2015).
8. Chris Mack, Fundamental Principles of Optical Lithography: the science of microfabrication, John Wiley & Sons (2008).
D
9. Alfred Kwok-Kit Wong, Resolution enhancement techniques in optical lithography,
PT E
Bellingham, Washington, SPIE press (2001). 10. Sascha Migura, Bernhard Kneer, Jens Timo Neumann, Winfried Kaiser, Jan van Schoot, “EUV lithography optics for sub 9 nm resolution”, International Symposium on Extreme
CE
Ultraviolet Lithography, Washington D.C., U.S.A. (2014). 11. Jan van Schoot, Koen van Ingen Schenau, Chris Valentin, and Sascha Migura, “EUV
AC
lithography scanner for sub-8nm resolution”, Proc. of SPIE, 9422, 94221F (2015). 12. Sudharshanan Raghunathan, Obert R. Wood, Pawitter Mangat, Erik Verduijn, Vicky Philipsen, Eric Hendrickx, Rik Jonckheere, Kenneth A. Goldberg, Markus P. Benk, Patrick Kearney, Zachary Levinson, and Bruce W. Smith, “Experimental measurements of telecentricity errors in high-numerical-aperture extreme ultraviolet mask images”, J. Vac. Sci. Technol. B, 32, 06F801 (2014). 13. Hoyoung Kang, Steve Hansen, Jan van Schoot, and Koen van Ingen Schenau, “EUV simulation extension study for mask shadowing effect and its correction”, Proc. SPIE, 6921, 69213I (2008)
ACCEPTED MANUSCRIPT
14. In-Seon Kim, Michael Yeung, Eytan Barouch, and Hye-Keun Oh, "Impact of a deformed extreme ultraviolet pellicle in terms of the critical dimension uniformity", Journal of Micro/Nanolithography, MEMS, and MOEMS, 15(2), 021003, (2016). 15. Yashesh A. Shroff, Pei-Yang Yan, and Michael Goldstein, “Progress in high transmission pellicles for EUVL”, 2007 international symposium on extreme ultraviolet lithography, Sapporo, Japan (2007).
PT
16. B. L. Henke, E. M. Gullikson, and J. C. Davis, “X-ray interactions: photoabsorption, scattering, transmission, and reflection at E ¼ 50-30000 eV, Z ¼ 1-92”, At. Data Nucl. Data
RI
Tables 54(2), 181–342 (1993).
SC
17. Craig De Young, “EUV-Supporting Moore’s Law”, Jefferies 2014 Global TMT
AC
CE
PT E
D
MA
NU
Conference, Miami, U.S.A. (2014).
SC
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
Graphical abstract
ACCEPTED MANUSCRIPT
Highlights Image distortion caused by wrinkled EUV pellicle is calculated. Intensity distribution after passing pellicle is varied with illumination condition s. Impact of wrinkled pellicle in terms of CD non-uniformity is presented
AC
CE
PT E
D
MA
NU
SC
RI
PT
Allowable limit of wrinkle of EUV pellicle is suggested for N5 patterning.
In-Seon Kim, Guk-Jin Kim, Michael Yeung, Eytan Barouch, HyeKeun Oh PII: DOI: Reference:
S0167-9317(17)30040-0 doi: 10.1016/j.mee.2017.01.031 MEE 10458
To appear in:
Microelectronic Engineering
Received date: Revised date: Accepted date:
14 October 2016 19 January 2017 25 January 2017
Please cite this article as: In-Seon Kim, Guk-Jin Kim, Michael Yeung, Eytan Barouch, Hye-Keun Oh , Impact of transmission non-uniformity of a wrinkled EUV pellicle for N5 patterning under various illuminations. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Mee(2017), doi: 10.1016/j.mee.2017.01.031
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Impact of transmission non-uniformity of a wrinkled EUV pellicle for N5 patterning under various illuminations a
In-Seon Kim, a Guk-Jin Kim, b Michael Yeung, c Eytan Barouch, and a Hye-Keun Oh† Applied Physics Department, Hanyang University, Ansan, Korea. b Fastlitho, SanJose, CA, USA. c Boston University, Boston, MA, USA. Email: †[email protected]
RI
PT
a
Abstract
SC
To solve the defect problem during the exposure process of EUV lithography (EUVL), an EUV pellicle is suggested as a solution. Even though use of an EUV pellicle is considered an
NU
essential solution for defect control during the exposure process, it is not ready for application in real commercial processes due to how difficult it is to manufacture. Due to
MA
tight requirements for an EUV pellicle, flawless fabrication is impossible and deformations such as a wrinkle would result in serious patterning problems. These deformations lead to critical dimension (CD) non-uniformity due to the non-uniform transmission distribution
D
caused by a pellicle wrinkle. In this paper, we discuss the impact of transmission non-
PT E
uniformity caused by a wrinkled pellicle. When we treat the effects of a wrinkled pellicle, we considered off-axis-illumination, which is a promising resolution enhancement technology. By shrinking the target pattern size down to 7 nm nodes (N7) or 5 nm nodes (N5), a flexible
CE
illumination system is required to enhance the resolution ability. With dipole illuminations, resolution is varied by partial coherence and by not only the beam size at the pellicle, but also
AC
by the incident angle distribution. For these reasons, a non-uniform intensity distribution caused by a wrinkled pellicle is modified with changes in off-axis illumination conditions, including spatial coherence and the aperture shape. For N5 patterning, the CD non-uniformity of 0.2 nm occurs at 1.7% transmission variation (2-pass). However, the intensity distribution is varied under various illumination conditions in spite of the same pellicle conditions. The intensity non-uniformity goes up with increasing spatial coherence. Even though feature conditions of the wrinkled pellicle are the same, the transmission non-uniformity is changed with illumination conditions. Thus, the allowable feature conditions of the wrinkled pellicle may be changed. For achieving good CD uniformity, the allowable limit of the pellicle
ACCEPTED MANUSCRIPT
wrinkle is carefully studied because the impact of the wrinkled pellicle varies with the illumination conditions.
AC
CE
PT E
D
MA
NU
SC
RI
PT
Keywords: EUV lithography; EUV pellicle; pellicle; pellicle deformation; Transmission non-uniformity; CD uniformity
ACCEPTED MANUSCRIPT
1. INTRODUCTION EUV lithography is one of the more promising technologies used to fabricate integrated chips with 1X nm dimensions and below [1, 2]. EUV lithography has an exceptional capability to produce sophisticated logic patterns because of the high resolution of UV light. However, some obstacles still need to be solved to allow for high volume manufacturing
[3]
. Among
various problems, defect control is a serious problem that must be solved before mass
pellicle is suggested as a solution
[2-4]
PT
production can take place. To solve the defect problem during the exposure process, an EUV . The EUV pellicle is a thin membrane that protects the
RI
mask from contamination by wrapping the mask. Even though the EUV pellicle is considered
SC
as an essential solution for defect control during exposure process, it is not ready for application in a real-world process because it is hard to precisely manufacture. Due to tight
NU
requirements for the EUV pellicle [5], flawless fabrication is impossible and it is very fragile. Deformations, such as a wrinkle, would cause a non-uniform transmission distribution involving the critical dimension (CD) [5-7]. CD non-uniformity is a critical problem for device
MA
production, thus the transmission non-uniformity study of a wrinkled pellicle is very important to realizing practical applications of the pellicle. By shrinking the target pattern size down to 5 nm nodes (N5), a conventional illumination [2, 8-10]
. In order to extend the resolution, we need a high NA system because it is
PT E
criterion
D
system with a 0.33 numerical Aperture (NA) meets the resolution limit based on Rayleigh
impossible to develop N5 pattering with a 0.33 NA system owing to the resolution limitation. For N5 patterning, a high 0.55 NA system was suggested, with a large magnification in order
CE
to avoid a serious shadow effect
[8-13]
. In spite of the high NA system, off-axis illumination
(OAI) technology is also required to enhance the resolution ability
[8-11]
. For large patterns,
AC
circular shaped illumination is usually used, but dipole illumination with small spatial coherence can achieve high contrast with fine patterns. Off-axis illumination technology can enhance the image quality of patterning but it can also affect wrinkled pellicle imaging. With dipole illumination, two poles of illumination are incident on the mask separately and they have a different incident angle distribution. Due to the incident angle distribution, the illumination source size at the pellicle level is varied. Therefore, the non-uniform intensity distribution caused by a wrinkled pellicle changes with variation in off-axis illumination conditions such as the aperture shape and spatial coherence. In this paper, we discuss the wrinkled pellicle effect for N5 pattering. In order to understand the influence of the wrinkled
ACCEPTED MANUSCRIPT
pellicle for the N5 pattering environment, we will treat not only the various illumination conditions but also various pattern sizes.
2. OPTICAL MODELING FOR WRINKLED PELLICLE WITH OAI As mentioned in the introduction section, a pellicle is easily deformed due to its weak structure and such deformation can lead to transmission non-uniformity. In order to calculate the image distortion caused by a wrinkled pellicle, we will optically model the wrinkled
PT
pellicle for calculation of transmission variation. First, we express the wrinkled depth of pellicle as a sinusoidal function, as shown in the schematic diagram of optical modeling in
PT E
D
MA
NU
SC
RI
Fig. 1 [14].
CE
Figure 1. Schematic diagram of the optical modeling for a wrinkled pellicle.
Height of wrinkled pellicle:
Eq. 1
AC
Since we can express the wrinkle height as a cosine function, we can calculate other feature conditions of the wrinkled pellicle. By differentiating the height function, we obtain a local slope at each position that represents the tangent value of the local tilt angle. Therefore, the arc tangent value of the local slope is a local incident angle, as shown in Eq. 2. Local tilt angle of wrinkled pellicle:
Eq. 2
With Eqs. 1 and 2, we can express the local tilt angle distribution as functions of the period and amplitude of the wrinkle. To calculate the optical path at the pellicle, we should know the incident angle for each position. The actual incident angle at the pellicle level is the sum of the local tilt angle of wrinkle and the angle distribution caused by OAI. If we use OAI technology, the incident angle is varied with aperture conditions. As shown in Fig. 2, EUV is
ACCEPTED MANUSCRIPT
obliquely incident at a 6 degree angle due to reflective optics and is focused at the mask and wafer
[2, 10]
. The incident angle distribution at mask level varies with spatial coherence, as
shown in Fig. 2(b). With 6 degrees of oblique incidence, the incident angle distribution caused by OAI can be expressed as Eq. 3. When we express the incident angle distribution, .
MA
NU
SC
RI
PT
we consider dipole illumination with spatial coherences (
AC
CE
PT E
D
(a)
(b) Figure 2. Light path for EUV lithography:
(a) EUV light path at projection optics and (b) illumination cone with dipole aperture at mask level.
Incident angle of Pole A: Incident angle of Pole B: The actual incident angle (
Eq. 3
at the wrinkled pellicle is a sum of Eqs. 2 and 3. With the
actual incident angle, the optical path at the pellicle (
can be calculated, as shown in Fig.
1 and Eq. 4. The absorption is exponentially proportional to the optical path and one pass transmission is presented in Eq. 4.
ACCEPTED MANUSCRIPT
one pass transmission of the wrinkled pellicle: Eq. 4 In Eq. 4, k is the extinction coefficient of the pellicle, dT is the optical path at the pellicle (shown in Fig. 1), λ is the wavelength of light, do is the thickness of the pellicle and θi is the actual incident angle. With this equation, we calculate the intensity distribution after passing
PT
through the pellicle once. As represented in Fig. 2(a), EUV light passes through the pellicle twice, and the beam position at the pellicle level is shifted due to the reflective optics and
RI
oblique incidence. Therefore, we should consider the shift caused by oblique incidence when we calculate the intensity distribution after passing through the pellicle twice. With
pellicle, which will be discussed in the next section.
SC
consideration of oblique incidence, we calculate the image distortion caused by the wrinkled
NU
3. CHARACTERISTICS OF TRANSMISSION NON-UNIFORMITY In this section, we will discuss image distortion caused by the wrinkled pellicle under
MA
various illumination conditions. When we calculate the transmission non-uniformity with the wrinkled pellicle, we consider the ideal case of the EUV pellicle. The ideal EUV pellicle was suggested by Intel first
[15]
. The ideal pellicle has a very simple structure,
D
which can achieve 90% transmission. The specifications of the ideal EUV pellicle are
PT E
summarized in Table 1. During the past few years, many different pellicle structures have been presented, but research regarding an optimized structure is still ongoing. Therefore, we use the ideal case at this time.
CE
Table 1. Simulation conditions of the EUV pellicle
c-Si 50 0.9991+0.0018i
AC
Material Thickness (nm) Refractive index (n+ik)[16]
3.1. Periodicity of transmission non-uniformity caused by oblique reflection. When the EUV light passes through the wrinkled pellicle, the incident angle varies with position. Due to the variation of incident angle, the optical path is also varied. By increasing the incident angle, the optical path is increased and thus causes a larger transmission loss, as shown in Fig. 3. As a result, the absorption is low at the top and bottom of a wrinkle, while the absorption at the sides of wrinkle is relatively higher due to a larger tilt angle and longer optical path.
PT
ACCEPTED MANUSCRIPT
RI
Figure 3. Incident angle and transmission variation of wrinkled pellicle.
SC
Owing to the transmission variation of a wrinkled pellicle, a wrinkled pellicle causes a non-uniform intensity distribution. One-pass transmission variation becomes very serious with a large tilt angle for the wrinkle. However, 2-pass transmission variation can be varied
NU
with period in spite of the same maximum local tilt angle because EUV light passes through the pellicle twice and incident light and reflected light pass at different positions due to
MA
oblique reflection. By considering transmission variation and passing position shift, we can
AC
CE
PT E
D
predict the cases of largest and smallest transmission difference, as shown in Fig. 4.
(a)
RI
PT
ACCEPTED MANUSCRIPT
PT E
D
MA
NU
SC
(b)
(c)
Figure 4. Periodicity of transmission non-uniformity of a wrinkled pellicle: (a) Conditions of the largest and smallest transmission difference cases, (b) intensity distribution
CE
after passing the pellicle twice, and (c) transmission non-uniformity as a function of the period of the wrinkle with the same maximum local tilt angle.
AC
In Fig. 4, graphs (a) and (b) show two cases of transmission variation. EUV lithography uses reflective optics with oblique incidence. Reflective optics and oblique incidence cause beam shift, as mentioned in the previous section. When the EUV light goes from a dark region to another dark region, the 2-pass transmission is a minimum. On the contrary, if the EUV light goes from a bright region to a bright region, the transmission is a maximum. On the left hand side of Fig. 4(a), the EUV light goes from a dark region to another dark region and a bright region to another bright region. Thus, the transmission difference will be large for this case. In contrast, when the EUV light goes from the bright region to dark region as in the case on the right hand side, the transmission difference is smaller than that of the left hand side case. As a result, the intensity distribution after passing through the pellicle twice
ACCEPTED MANUSCRIPT
is different for the two cases, as shown in Fig. 4(b). This indicates that the transmission nonuniformity is different with a different period of the wrinkle, despite having the same maximum local tilt angle. Figure 4(c) shows transmission non-uniformity as a function of the wrinkle period with a 300 mrad maximum local tilt angle. In Fig. 4(c), it is evident that transmission nonuniformity fluctuates with the period of the wrinkle. When the beam shift due to oblique
PT
reflection is a multiple of a half-period of the wrinkle, the transmitted light non-uniformity is at its largest. However, when the shift is a multiple of a half-period plus or minus a
RI
quarter period due to a different period, the 2-pass intensity non-uniformity is at its smallest.
SC
With consideration of the shift due to oblique incidence, the last case of periodicity is a wrinkle with a 1.05 mm period. Therefore, if the period is larger than about 1 mm, the periodicity of transmission non-uniformity does not appear.
NU
3.2. Beam size effect of partial coherent light.
Until now, we have explored pellicle imaging variation caused by reflective optics. Now,
MA
we will discuss the wrinkled pellicle imaging variation with partial coherent light. For resolution enhancement, various illumination aperture shapes and spatial coherences are used. By increasing the partial coherence, the beam size of the illumination cone at
D
pellicle will be increased. Owing to the variation of beam size, wrinkled pellicle imaging
PT E
can be varied. With dipole illumination, the intensity distribution after passing the pellicle
AC
CE
twice is shown in Fig. 5.
Figure 5. Wrinkled pellicle imaging with various illumination conditions.
ACCEPTED MANUSCRIPT
In Fig. 5, results show the intensity distribution after passing through the pellicle twice with various dipole sources. The illumination aperture shape is a dipole and the outer sigma is 0.9. As shown in Fig. 5, the intensity distribution caused by a wrinkled pellicle is varied with spatial coherence and the large transmission non-uniformity with a small dipole source (σin=0.8) is significantly decreased with a large dipole source (σin=0.2). In order to understand the beam size effect of a wrinkled pellicle, we separated the range of
PT
the wrinkle period. Figure 6 shows a 0.55 NA system at mask level. The suggested 0.55 NA system has an 8X magnification and the stand-off height between the mask and pellicle is
RI
fixed as 2.5 mm [5, 10]. With consideration of this structure, the maximum beam size with the
SC
0.55 NA system is about 350 μm. If the period of the wrinkle is larger than 1 mm, the distance between the dark region and bright region is far compared with the maximum beam size. As a result, the transmission non-uniformity caused by a wrinkled pellicle is decreased
PT E
D
MA
NU
in the large wrinkle region, as shown in Fig. 7.
AC
CE
Figure 6. Illumination system at the mask level with a 0.55 NA system.
ACCEPTED MANUSCRIPT
(b)
SC
RI
PT
(a)
Figure 7. Beam size effect of OAI with a wrinkled pellicle:
NU
(a) Transmission non-uniformity variation with various periods of wrinkle. (b) Transmission non-uniformity difference between a small and large wrinkle.
MA
Figure 6 shows the beam size effect of OAI with a wrinkled pellicle. If the period of the wrinkle is small, the transmission non-uniformity varies with aperture conditions due to the beam size effect. However, the difference of transmission non-uniformity with illumination
D
conditions is decreased as the period of the wrinkle is increased. Thus, if the period of the
PT E
wrinkle is much larger than the maximum beam size, the transmission non-uniformity is independent of aperture conditions. As can be seen in Fig. 6(b), the transmission nonuniformity as a function of the maximum local tilt angle also varies with illumination
CE
conditions in the small wrinkle region, but it is independent of illumination conditions in the large wrinkle region. With this criterion, the CD non-uniformity caused by the wrinkled
AC
pellicle may differ with illumination conditions in the small verses large wrinkle regions. This will be discussed in the next section.
4. CD NON-UNIFORMITY CAUSED BY A WRINKLE PELLICLE WITH N5 PATTERNS. After the EUV light passes through the pellicle twice, the wrinkled pellicle leads to a nonuniform intensity distribution, which causes CD non-uniformity. The CD difference between the bright region and dark region is linearly proportional to the intensity difference [14]
. In this section, we will measure the CD non-uniformity as a function of transmission
non-uniformity and then discuss the CD non-uniformity of N5 patterns caused by a wrinkled pellicle.
ACCEPTED MANUSCRIPT
In order to know the impact of the wrinkled pellicle for N5 patterns, we consider an isoline with a dense period pattern and N5 specifications, and measure CD variation as a function of transmission non-uniformity with 2 cases of dipole illumination conditions. Table 2. Simulation conditions for the CD non-uniformity characteristics study [16, 17].
EUVL suite (Fastlitho Inc.)
Simulation tool Iso-line
5
(nm)
Dense line & space
11 − 15
Tantalum boron nitride (TaBN)
Thickness
70 nm
Refractive index (n+ik) Small dipole
conditions
Large dipole
NU
Illumination
0.9165 + 0.0438i
SC
conditions
Material
RI
Absorber
PT
Target CD
The simulation conditions are summarized in Table. 2. With these conditions, we measured
MA
CD variation with a non-uniform intensity distribution and CD non-uniformity as a function of 2-pass transmission non-uniformity, as depicted in Fig. 8. In Fig. 8, graph (a) is the result
AC
CE
PT E
D
with the large dipole and graph (b) is the result with the small dipole.
(a)
RI
PT
ACCEPTED MANUSCRIPT
(b)
SC
Figure 8. CD non-uniformity as a function of transmission non-uniformity for N5 patterns.
If we define the slope value as a sensitivity, sensitivities are different not only with respect
CD uniformity budget is under 0.2 nm
[5]
NU
to pattern size, but also with respect to illumination conditions. For the EUV pellicle, the . To obtain 0.2 nm CD uniformity, the 2-pass
transmission non-uniformity should be smaller than 1.7% for all patterns, regardless of
MA
illumination conditions. The allowable transmission non-uniformity for N5 patterns is the same as for two dipole illuminations, but the intensity distribution caused by a wrinkled
D
pellicle varies with illumination conditions, as mentioned in the previous section. The CD non-uniformity as a function of maximum local tilt angle of the wrinkle is summarized in
AC
CE
PT E
Fig. 9.
Figure 9. CD non-uniformity as a function of the maximum local tilt angle of the wrinkle with dipole illumination.
In Fig. 9, gray lines are the results of the large dipole and black lines are that of the small dipole. The CD non-uniformity results in the small wrinkle region are expressed with solid lines and results in the large wrinkle region are described with dashed lines. As mentioned
ACCEPTED MANUSCRIPT
in Section 3, the transmission non-uniformity decreases with large spatial coherence in the small wrinkle region. As a result, the CD non-uniformity of the large dipole is smaller than that of the small dipole in the small wrinkle region. To obtain 0.2 nm CD uniformity, the maximum local tilt angle of the wrinkle should be smaller than 420 mrad with small dipole illumination for all N5 patterns. In the case of large dipole illumination, the CD uniformity variation is smaller than that of the small dipole and the requirement of maximum local tilt
PT
angle is relaxed to about 520 mrad. Thus, a large dipole gives a large allowable range of wrinkle compared to small dipole illuminations. In contrast to the small wrinkle, the
RI
transmission non-uniformity caused by a wrinkled pellicle is independent of the illumination
SC
conditions. Due to similar transmission variation and the same requirements of transmission non-uniformity, the requirement of maximum local tilt angle is almost the same for two dipole illuminations in the large wrinkle region, and the shared requirement of maximum
NU
local tilt angle is about 600 mrad.
5. CONCLUSION
MA
The EUV pellicle is an essential solution for the protection of a mask and it should have an extremely thin thickness for high transmission. Due to the weak structure of the pellicle, a wrinkled pellicle is an unavoidable problem. A pellicle wrinkle leads to a non-uniform
D
intensity distribution caused by transmission variations. Owing to the reflective optics and
PT E
oblique incidence of EUVL, the intensity non-uniformity varies according to the period of the wrinkle after passing the pellicle twice. With the consideration of oblique reflection and transmission distribution of the wrinkle, the last case of a large transmission difference is
CE
the wrinkle with a period of 1 mm. If the period of the wrinkle is larger than about 1 mm, the worst case of transmission non-uniformity is avoided completely. In the large wrinkle
AC
region, local intensity variation is relatively independent of not only the reflective optics, but also the illumination conditions of OAI. As the beam size varies at the pellicle, the intensity non-uniformity also varies, and it is increased with highly coherent light and a small wrinkle. However, the intensity non-uniformity is independent of illumination conditions in the large wrinkle region. With dipole illumination, the CD non-uniformity of 0.2 nm occurs with a transmission variation (2-pass) of 1.7%, regardless of illumination conditions. However, the allowable feature condition of a wrinkled pellicle is different due to different pellicle imaging results. For N5 patterning, the requirements of a maximum local tilt angle are 420 mrad for a small dipole and 520 mrad for a large dipole illumination.
ACCEPTED MANUSCRIPT
ACKNOWLEDGEMENT This research was supported partly by Samsung electronics mask development team and partly by the Nano Material Technology Development Program through the National Research Foundation of Korea (NRF), and is funded by the Ministry of Education, Science
AC
CE
PT E
D
MA
NU
SC
RI
PT
and Technology (No. 2015M3A7B7045353).
ACCEPTED MANUSCRIPT
REFERENCES 1. L. Wilson, “International Technology Roadmap for Semiconductors (ITRS)”, Semiconductor Industry Association (2013). 2. Vivek Bakshi, EUV Lithography, Bellingham, Washington, SPIE Press (2009). 3. Kurt Ronse, “Closing address”, International symposium on Extreme Ultraviolet lithography, Maastricht, Netherlands (2015).
PT
4. Harry J. Levinson, Principle of lithography, Bellingham, Washington, SPIE press (2005). 5. Derk Brouns, "Pellicle HVM specifications". Pellicle Technical Meeting Group, San Jose,
RI
USA. 2016.
SC
6. Luigi Scaccabarozzi, Dan Smith, Pedro Rizo Diago, Eric Casimiri, Nina Dziomkina, and Henk Meijer, “Investigation of EUV pellicle feasibility”, Proc. of SPIE, 8679, 867904 (2013).
NU
7. Dario L. Goldfarb, Max O. Bloomfield, and Matthew Colbum, “Thermo-mechanical behavior of EUV pellicle under dynamic exposure conditions”, Proc. of SPIE, 9776, 977621
MA
(2015).
8. Chris Mack, Fundamental Principles of Optical Lithography: the science of microfabrication, John Wiley & Sons (2008).
D
9. Alfred Kwok-Kit Wong, Resolution enhancement techniques in optical lithography,
PT E
Bellingham, Washington, SPIE press (2001). 10. Sascha Migura, Bernhard Kneer, Jens Timo Neumann, Winfried Kaiser, Jan van Schoot, “EUV lithography optics for sub 9 nm resolution”, International Symposium on Extreme
CE
Ultraviolet Lithography, Washington D.C., U.S.A. (2014). 11. Jan van Schoot, Koen van Ingen Schenau, Chris Valentin, and Sascha Migura, “EUV
AC
lithography scanner for sub-8nm resolution”, Proc. of SPIE, 9422, 94221F (2015). 12. Sudharshanan Raghunathan, Obert R. Wood, Pawitter Mangat, Erik Verduijn, Vicky Philipsen, Eric Hendrickx, Rik Jonckheere, Kenneth A. Goldberg, Markus P. Benk, Patrick Kearney, Zachary Levinson, and Bruce W. Smith, “Experimental measurements of telecentricity errors in high-numerical-aperture extreme ultraviolet mask images”, J. Vac. Sci. Technol. B, 32, 06F801 (2014). 13. Hoyoung Kang, Steve Hansen, Jan van Schoot, and Koen van Ingen Schenau, “EUV simulation extension study for mask shadowing effect and its correction”, Proc. SPIE, 6921, 69213I (2008)
ACCEPTED MANUSCRIPT
14. In-Seon Kim, Michael Yeung, Eytan Barouch, and Hye-Keun Oh, "Impact of a deformed extreme ultraviolet pellicle in terms of the critical dimension uniformity", Journal of Micro/Nanolithography, MEMS, and MOEMS, 15(2), 021003, (2016). 15. Yashesh A. Shroff, Pei-Yang Yan, and Michael Goldstein, “Progress in high transmission pellicles for EUVL”, 2007 international symposium on extreme ultraviolet lithography, Sapporo, Japan (2007).
PT
16. B. L. Henke, E. M. Gullikson, and J. C. Davis, “X-ray interactions: photoabsorption, scattering, transmission, and reflection at E ¼ 50-30000 eV, Z ¼ 1-92”, At. Data Nucl. Data
RI
Tables 54(2), 181–342 (1993).
SC
17. Craig De Young, “EUV-Supporting Moore’s Law”, Jefferies 2014 Global TMT
AC
CE
PT E
D
MA
NU
Conference, Miami, U.S.A. (2014).
SC
RI
PT
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
Graphical abstract
ACCEPTED MANUSCRIPT
Highlights Image distortion caused by wrinkled EUV pellicle is calculated. Intensity distribution after passing pellicle is varied with illumination condition s. Impact of wrinkled pellicle in terms of CD non-uniformity is presented
AC
CE
PT E
D
MA
NU
SC
RI
PT
Allowable limit of wrinkle of EUV pellicle is suggested for N5 patterning.