Graphite Pellicle: Physical Shield for Next‐Generation EUV Lithography Technology

Extreme ultraviolet lithography (EUVL) is widely employed in the electronics, automotive, military, and AI computing areas for IC chip fabrication. A pellicle is a thin and transparent membrane that protects a costly photomask, known as a reticle, during the EUVL process. The fabricated IC chip can be disastrous without a pellicle. When a particle lands on a photomask, it frequently results in a faulty pattern, which leads to chip failure and lower production yield. A nanometer‐thick graphite (NGF) has demonstrated tremendous potential for addressing optical, mechanical, thermal, and chemical criteria among potential pellicle materials such as carbon allotropes, Si, SiNx, and Si‐Mo‐Nb. This review summarizes current progress in NGF pellicles, including large‐scale material fabrication (up to 135 mm × 135 mm), transfer method for freestanding form, and practical characterization methods. Current significant challenges and future opportunities for NGF pellicles are also discussed in order to facilitate a critical transition from lab‐scale research to industrial‐scale implementation.


Introduction
The extreme ultraviolet (EUV) pellicle as a physical shield for the photomask is emerging and becoming essential for the fabrication of nanodevice under 5 nm. The microparticles inevitably exist in the EUV lithography (EUVL) chamber with a hydrogen environment (3 Pa), which can have an methods are summarized that verify NGF properties as a pellicle.

Graphene/Graphite Synthesis for Pellicle Application
Graphene, which is a single atomic layer containing only carbon atoms with sp 2 hybridized chemical bonds, possesses outstanding thermal and electrical conductivity, chemical and mechanical properties, and a protective barrier to isolate metal surfaces, allowing it to be implemented as a pellicle that meets key requirements [1,2] for EUVL pellicle technology. The first step in the growth of large-scale single-layer graphene on Cu foil and multilayered graphene on Ni film/ silicon substrate by CVD technique was to verify the synthetic mechanisms on the self-limiting way [8] in 2009 and carbon solubility of Ni/ carbon precipitation during cooling [9] in 2009, respectively. In the beginning, graphene was mainly utilized for the transparent electrode [10] and graphene field-effect transistor, [11] but there was no technical interest in the pellicle application in EUVL. Furthermore, despite its exceptional strength of 130 GPa (within 1.5 µm diameter of graphene film only), [12] single-layer graphene was too thin to be freestanding on a frame with the whole size of pellicle 110 mm × 140 mm in air. In 2015, the NGF pellicle with 50 mm × 50 mm size was demonstrated for the very first time. [2] At first, the critical issues for realizing NGF pellicle were i) uniform NGF synthesis with lower defect density, ii) conformal NGF transfer on the frame, and iii) large-scale characterization tool development. In this section, we will review the developing synthetic method for NGF pellicles.

NGF Synthesis on the Cu Foil
As mentioned above, graphene and multilayer graphene can be synthesized using the CVD technique with Cu or Ni metal catalyst. In particular, NGF (10-100 nm) can only be obtained through a specific combination of hydrocarbon and metal catalyst, e.g., acetylene-Cu and methane-Ni (Figure 2). Multilayered graphene could be synthesized with methane-Cu, although single-layer graphene (>95%) is dominantly grown on the surface due to its self-limiting behavior (or surface reaction). [8] The presence of the first layer is necessary for implementing lateral growth of another layer on it. Even with a high hydrocarbon (reactive gas) concentration and long growth time, it is rare to grow bi-(3%-4%) and tri-layer (<1%) graphene on large scale. [13] Similarly, it is found that the Cu grain orientation of polycrystalline foil has little or no effect on the presence/formation of multilayer regions (Figure 3a). [14] In the case of acetylene-Cu, there is a greater chance of surface interaction during the synthesis of bilayer or multilayered graphene because decomposing hydrocarbons at the surface of Cu produce more active carbon species depending on the H 2 /Ar ratio (Figure 3b). [15] Despite the synthetic strategies developed for the graphene and a few layers of graphene on the Cu foil by CVD, which are suitable for electronic device applications, they lag behind for the EUV pellicle application.
We reported the synthesis of NGF using the combination of acetylene and Cu for pellicle applications in 2017. [16] Tens of nanometers of NGF (from 1.5 to 65 nm) were realized by controlling the C 2 H 2 flow rate (20-300 sccm) and growth time (1-8 h) (Figure 3c). To measure the thickness of NGF using AFM, i) the synthesized NGF on Cu foil was first coated with poly(methyl methacrylate) (PMMA) supporting layer, ii) the Cu www.advmatinterfaces.de was etched away by the FeCl 3 solution, iii) NGF/PMMA was cleaned with HCl and water several times, and iv) it was transferred onto a Si substrate.

NGF Synthesis on the Ni Foil
Among the transition metals, Ni has been proposed as a catalyst for graphene synthesis. Carbon solubility primarily determines whether graphene forms at the surface of the metal at high growth temperatures, as is the case for Cu, or whether carbon dissolves into the bulk and precipitates from the bulk, as is the case for Ni. [17,18] In 2009, the first step in the growth of a few layers of graphene on the Ni foil using the CVD technique was suggested including the pre-annealing in a H 2 environment at 1000 °C, introducing the reactive gas mixture (C 2 H 2 /H 2 ), and cooling down to 500 °C (160 °C min −1 ). [19] The dark-gray color started being observed at a growth time of 5 min, which was identified as the formation of a few graphene layers. As the growth time increased, a few white spots were filled with additional graphene layers (Figure 4a). In 2015, it was demonstrated for the first time that a 100 mm × 100 mm size of NGF can be grown using a rapid temperature annealing process for EUV pellicle applications. [2] Particularly, NGFs with various thicknesses were synthesized by controlling the growth temperature ranging from 910 to 1035 °C (910, 925, 1000 °C in Figure 4b). We observed well-arranged layer structure and layer distances that are very close to the theoretical layer distance of the graphite,  . Graphene/graphite synthesis on Cu foil. Depending on reactive gases, such as i) methane: a) control of size, shape, and thickness of graphene domains by pre-annealing of Cu; Reproduced with permission. [14] Copyright 2018, Elsevier B.V.; and ii) acetylene: b) high-quality synthesis of bilayer graphene by the pre-annealing of Cu with H 2 /Ar; Reproduced with permission. [15] Copyright 2013, American Chemical Society; c) NGF synthesis optimized by C 2 H 2 flow and growth time; Adapted with permission. [16] Copyright 2017, Elsevier B.V. www.advmatinterfaces.de 0.335 nm in high-resolution TEM cross-section images. At the beginning of the synthesis of NGF, it was not possible to form large-area NGF (over 100 mm × 100 mm) with a thickness of less than 70 nm by using standard CVD techniques (one-step) consisting of three processes including pre-annealing, growth, and cooling, as depicted in Figure 4c. Therefore, 80-100 nm thick NGF was initially synthesized on Ni foil, transferred with a PMMA supporting layer onto a frame, and then etched to ≈60 nm thick using O 2 plasma. [2,20] After etching PMMA by O 2 plasma, the NGF can be inevitably damaged, as evidenced by enhanced D-band intensity in the Raman spectrum due to the increment of defect density on the basal plane of NGF. To solve that issue, a new CVD technique, a two-step CVD process that adds a cooling step in the middle of the growth and cooling steps, was proposed. [20] From this technique, the large-scale (120 mm × 120 mm) NGF with a thickness of 38 nm were directly synthesized at a relatively low growth temperature (870 °C). Still, the thickness of grown NGF is required to be less than 20 nm without any etching procedures if EUV pellicle transmission is to surpass 90%. For the enhancement of NGF uniformity with a thickness of around 20 nm, a CVD technique involving a rapid cooling step and a separate precipitation step before the cooling step was proposed. [21] As long as the increased cooling rate to room temperature (RT) is sufficiently fast (>20 °C s −1 , c.f. conven-tional cooling rate: 9 °C s −1 ), the carbon atoms are frozen in the Ni, the supersaturated solid solution is maintained, and graphite does not form on the Ni surface at RT. In the separated precipitation step at the relatively low temperature (700 °C), high-density and homogeneous nuclei are formed on the Ni surface without the sparsely grown graphene/graphite nuclei at 910 °C, resulting in improved thickness uniformity and coverage of the NGF (Figure 4d).

NGF Synthesis on the Ni Film
In general, structural flaws in Ni foil impede large-scale NGF growth by CVD, including polycrystalline grain size, microbumpy (straight line shape) presented over the entire foil area, and flutter and folding of thin foil in preparation for CVD growth, which causes the structural defects of NGF such as wrinkles, cracks, nonuniformity, etc. Therefore, the ultraflat, robust growth technique of NGF becomes critical for the EUV pellicle. To this end, the strategies such as the deposition of Ni film on silicon substrates and controlling crystalline grain and surface roughness have been developed for large-scale NGF growth with tens of nanometers thick.
The first prototype of few-layer graphene on a Ni film was made in 2009, to develop a transparent electrode for a touch  [19] Copyright 2009, Wiley-VCH GmbH, Weinheim. b) TEM images of NGF grown at 910, 925, and 1000 °C. Reproduced with permission. [2] Copyright 2015, The Royal Society of Chemistry. c) Two-step CVD process for the growth of NGF with a thickness of 38 nm over large scale. Reproduced with permission. [20] Copyright 2017, Elsevier B.V. d) The CVD process with an additional fast cooling and a separated precipitation steps before the cooling step for the NGF growth for less than 20 nm thickness. Reproduced with permission. [21] Copyright 2018, Elsevier B.V.

www.advmatinterfaces.de
sensor. [9] After 10 years, a hybrid metal substrate comprised of extra Ni film on Ni foil was proposed for large-scale NGF growth showing improved uniformity. The optical image of raw Ni foil in Figure 5a shows a large polycrystalline grain size (average 32 µm) and microbumpy. [22] Essentially, carbon precipitation in Ni during the cooling step begins along the Ni grain boundaries, resulting in thicker NGF on those line defects. [20] The nonuniform thickness of the NGF pellicle changes the pellicle's EUV transmittance, causing an interruption of the fidelity pattern on the wafer during EUVL. The pellicle is at an out-of-focus position on the EUV imaging path of the scanner (Figure 1), allowing the pellicle's minimum transmission uniformity measurement to be tens of microns. [23] If the Ni grains are small enough, the influence on the formation of thicker NGF at Ni grain boundaries is averaged out over a broader EUV beam size, which can be considered relatively uniform. [22] The optimum thickness of Ni film sputtered on Ni foil is 1 µm, resulting in an average grain size of ≈6 µm. After a two-stage CVD process, an NGF pellicle with 86.3 ± 0.9% was achieved (c.f. 79 ± 2% for NGF grown on Ni foil).
As illustrated in Figure 5b, based on the report on the hybrid Ni substrate, a Ni film directly sputtered on the SiO 2 /Si substrate (40 mm × 40 mm) was adopted for NGF growth (20 nm thickness, 87.5% of EUV transmittance). [24] Subsequently, we recently developed an NGF pellicle of 100 mm × 100 mm, which is realized by the successful NGF growth on Ni film/ SiO 2 /Si substrate (135 mm × 135 mm) and camphor transfer method. [25] For the scalability, a 1.5 µm thick Ni film was sputtered on an 8-in. SiO 2 /Si wafer, then NGF was synthesized using a homemade cold-wall CVD system ( Figure 5c). The chamber was heated up to 900 °C under H 2 environment, then CH 4 gas was introduced into the chamber and the temperature was held at 900 °C for 10 min. After 10 min, the chamber was cooled down to 700 °C to grow the large-scale NGF with a thickness of ≈20 nm (87.2 ± 0.78% at 30 mm × 30 mm size of NGF pellicle).

Transfer Methods for Freestanding NGF Pellicle
Graphene transfer can be traced back to the conventional mechanical exfoliation process, in which graphene flakes were simultaneously exfoliated and transferred from graphite to target surfaces. Graphene/ NGF transfer from catalytic metals to a target substrate, in contrast to the conventional one, confronts greater hurdles due to the chemical etching of metal in a strong acid bath and transport of graphene between cleaning solutions, resulting in wrinkles, cracks, and fracture of a thin film. Thus, a layer, which can protect graphene/ NGF (supporting layer) from mechanical damage during the transfer process, such as PMMA, must be required. Furthermore, for large-scale pellicles, a freestanding criterion is essential. From a freestanding perspective, earlier works reported employing a TEM grid as a template frame while the single-layer graphene was supported by a Cu frame (1.3 µm size freestanding). [26][27][28] However, the transfer technique is not yet mature enough to realize the large-scale pellicle. The review article summarizes many NGF transfer processes for a freestanding criterion beyond graphene transfer methods developed so far, including laboratory measures scaled up to industry standards.
This study surveyed the etching and etching-free supporting layer after the transfer, as well as representative methods of supporting layers for enhancing the properties of graphene film after delamination from a sacrificial layer. Among the graphene transfer approaches for forming a pellicle, sacrificial substrate etching (PMMA by O 2 plasma or sublimation of camphor) has been the overall winner owing to its structural integrity and easy scalability. In addition, freestanding graphene pellicles have been realized without a transfer procedure depending on the metal species used for NGF synthesis by CVD. We compare and highlight the benefits and drawbacks of each technique by summarizing their underlying concepts. In this section, we limit the scope of NGF transfer advances for pellicle application in order to emphasize practical scalability. Reproduced with permission. [22] Copyright 2019, Elsevier B.V. b) NGF growth on Ni film (40 mm × 40 mm). Adapted with permission. [24] Copyright 2021, IOP Publishing Ltd. c) NGF growth on Ni film (135 mm × 135 mm). Reproduced with permission. [25] Copyright 2022, The Authors, published by the Royal Society of Chemistry. www.advmatinterfaces.de

Wet/Dry Transfer of NGF
As a proof-of-concept, we reported the wet and dry transfer (WaDT) method in 2015 for the fabrication of large-scale NGF pellicles, which had a freestanding size of 50 mm × 50 mm. [2] The WaDT procedure is depicted as a schematic in Figure 6. After etching the Ni substrate, NGF that had been coated with PMMA was subsequently placed onto a polycarbonate-tracketched (PCTE) membrane. PMMA/NGF/PCTE that had only partially dried were adhered to a frame using adhesive. PCTE was vital to the process of sustaining damp layers and aiding their transfer to the frame. The PCTE membrane was easily separated from the PMMA/NGF, and the PMMA was removed by O 2 plasma treatment. To make a large-scale NGF pellicle with high EUV transmission, we manipulated the thickness of NGF using O 2 plasma. The EUV transmission map indicated that the thickness of the freestanding NGF was 59.73 ± 5.83 nm.
Besides, we observed that the areal nonuniformity of EUV transmission was caused by significant wrinkles of freestanding NGF pellicle owing to the ambient condition process.

Vertical Transfer of NGF
ASML, the sole EUV tool supplier, has annually reported that the required size for pellicles is 132 mm × 104 mm till 2022 (effective sit area). [29] To meet this dimensional requirement, we reported a thorough investigation of the graphene-liquid interface during the NGF transfer process to fulfill the practical scalability of 100 mm × 100 mm and to acquire a deeper comprehension of the transfer process as shown in Figure 7. [30] Understanding the relationship between NGF, water, and air allowed us to characterize the underlying principle of the fabrication method. We captured the interaction in high-speed www.advmatinterfaces.de optical photography at several transferring angles to identify the total Gibbs free energy during the transfer process. By using the vertical transfer approach, we were able to successfully build a 100 mm × 100 mm freestanding pellicle. The configuration at three-phase interfaces, NGF, water, and air, altered from the reference state to a given transfer angle was explored based on Neumann and Good theory with interface imaging. The least change in total Gibbs free energy was found at a transfer angle of 90°, indicating that NGF film is more thermodynamically stable than other angles, leading to experimental validation. Using the vertical transfer technique, the NGF was moved from a floating state over deionized water to a square frame (FST, aluminum, 1 mm thick) with an inner hole measuring 100 mm × 100 mm. When the PMMA was dry, it was removed using a 10-minute O 2 plasma treatment (100 W, 100 sccm). This study disclosed the interaction of NGF with water and air, resulting in a feasible approach to NGF pellicle manufacturing employing a specific liquid phase and transfer angle.

Alternative to PMMA Supporting Layer
More subsequent studies have discussed possible alternatives to PMMA as PMMA-assisted transfer process has shown several drawbacks. Even after a thorough cleaning with acetone, PMMA residues often persist over a considerable area. Because of the robust contacts between PMMA and the graphene layer, structural flaws such as folding and ripping occur during the etching and transfer processes. PMMA removal with acetone or O 2 plasma, performed after transferring the NGF onto the frame, resulted in issues including residues and NGF deflection caused by acetone's high surface tension and was therefore deemed unnecessary. This was observed throughout the pellicle fabrication process. Similar to wet removal of PMMA, O 2 plasma removal caused abnormalities of NGF including deflection and defects.
For these reasons, camphor as an alternative to PMMA for use as a supporting layer was reported (Figure 8). A camphor Figure 7. NGF pellicle fabricated by vertical transfer (VT) process. a) Schematic illustration of wet etching of Ni foil underlying graphite film and VT process. b) Semitransparent 100 mm × 100 mm graphite pellicle. Scheme of the configurations at the interfaces with graphite film, water, and air at VT angles of c) at 90° and d) below 90° (variables in the theory of Neumann and Good to calculate the total Gibbs free energy). e) High-speed photo images at the interface between graphite film, water, and air to extract the equilibrium contact angle. Reproduced with permission. [30] Copyright 2020, Wiley-VCH GmbH.
www.advmatinterfaces.de film was deposited on the NGF layer using vacuum evaporation and served as a supporting layer during the transfer process. As the camphor has shown promising performance in smallarea graphene transfer (Figure 8a,b), subliming or removing a camphor layer was straightforward in either an air or ethanol environment.
We investigated the conditions of camphor deposition/ removal and optimized for large-scale pellicle fabrication either in wet or dry transfer processes and compared them to those with the PMMA transfer process. [24,25] In contrast to the traditional practice of using a PMMA supporting layer, camphor sublimates at room temperature, thus it does not leave any residues on the NGF. Compared to other supporting layer materials including SPPOI, [31] rosin, [32] pentacene, [33] and PMMA, [34] camphor has the lowest adsorption energy of 0.09 eV to graphene, which is the reason for the selection of camphor.
However, it was pointed out that the thickness of camphor as a function of temperature and pressure is a key parameter due to the complete sublimation during the transfer process. The standard method for realizing a freestanding graphene/ graphite pellicle, transferred PMMA/NGF film on DI water (rinsing purpose after metal substrate etching), experiences high surface tension leading to wrinkling and deflection of NGF film. A stiff camphor supporting layer (140 µm thick), on the other hand, was used to fabricate a freestanding NGF pellicle. Using a confocal displacement sensor with a resolution of 260 nm, the deflection of the transferred NGF (12.6 ± 2.2 µm) was ten times less than that of the NGF using PMMA (127 ± 28.9 µm). To validate the use of camphor without residue after sublimation, Raman spectrum exhibiting only graphene peak and showing no difference in D/G peak intensity ratio (I D /I G ) was identified between as-grown graphite film and freestanding NGF pellicle. A comparable study employing camphor as a supporting layer was reported with a focus on the chemical stability of a pellicle against hydrogen plasma. [35]

Transfer-Free Pellicle Fabrication
Transfer-free, freestanding pellicle fabrication was demonstrated as a proof-of-concept and schematically shown in Figure 9. [16] NGF film was grown on Cu foil without the need for a transfer step during the entire pellicle fabrication. Similar to the wet-etching of Ni or Cu foil, PMMA was used to protect the NGF layer. Uniquely, PDMS was selectively deposited along the border of the Cu foil (a-1) where the area of uncoated Cu foil was removed by dipping in FeCl 3 solution (a-2). PMMA/ NGF/Cu/PDMS was cleaned by switching solvents from HNO 3 to DI water (a-3), followed by acetone (removal of PMMA, a-4) resulting in a freestanding NGF pellicle (drying, a-5). All the solvents used during the etching, cleaning, and drying processes should be switched/drained by a peristaltic pump at each step in order to diminish mechanical damage caused by surface tension differences on the freestanding NGF. [16] This study demonstrated the feasibility of fabricating a large-area freestanding pellicle without a transfer procedure as well as satisfying transmittance of over 93% (Figure 9b,c). However, it was clearly stated that some dark areas (thick graphite film parts) or Cu residues caused nonuniformity implying that further study is required.
We have surveyed transfer techniques for fabricating freestanding pellicles with an emphasis on large-scale NGF. Table 1 summarizes the reported progress in the NGF pellicles and their critical parameters for a transfer method.

Characterization Methods for Freestanding NGF Pellicle
In previous sections, the strict requirements of pellicle from a material standpoint were outlined, including EUV transmission of more than 90%, excellent thermal and chemical stability upon exposure to EUV (pulse-scan conditions of 188 ms exposure and 100 ms break [1] ), mechanical endurance to frame manipulating speed, maximum ambient pressure rate of change, and transfer strategies for satisfying dimensional requirement (practical size of 110 mm × 140 mm). The crystallinity of NGF, its mechanical, thermal, and chemical properties, as well as morphology and deformation before and after the transfer to the frame must be established in order to demonstrate the viability of the NGF pellicle. In this section, essential characterization methods for freestanding NGF pellicles are summarized.

Crystallinity of NGF
We demonstrated the crystallinity of NGF by Raman spectroscopy, TEM, XRD, etc., because point defects (voids, pinholes) and line defects (grain boundaries) could lead to severe degradations of NGF pellicle. In particular, Raman spectroscopy is a nondestructive chemical analysis technique and an effective measurement tool for detecting vibrational, rotational, and other states in a molecular system, which can provide chemical structure, phase and polymorphism, crystallinity, and molecular interactions. Outstanding results on doping, edges, strain and stress, disorder, oxidation, hydrogenation, chemical functionalization, electrical mobility, thermal conductivity, electron-phonon and electron-electron interaction, magnetic field, and interlayer coupling have contributed to a significant advancement in the understanding of Raman spectroscopy in the graphene research field. [36] Crystallinity, grain size, and relative defect density of NGF synthesized on Cu and Ni transition metal can be readily extracted due to the understanding of Raman spectroscopy for graphene using the intensity ratio of the D band to the G band (I D /I G ), the full width of half maximum of the 2D band (W 2D ), and the intensity ratio of the 2D band to the G band (I 2D /I G ). Figure 10 displays the Raman spectra of NGF synthesized on Cu foil, Ni foil, and Ni film (sputtered). NGF synthesized with acetylene gas on Cu foil by controlling the hydrocarbon gas flow rate and growth time [16] has a relatively high defect density from which the I D /I G ratio, 0.43 to 0.69, of Raman spectra of NGF with increasing growth time from 1 to 8 h (Figure 10a [24] Copyright 2021, IOP Publishing. c) Photographs of the mechanical peeling-off process of the NGF using temporary transfer film (TTF), d) photographs of NGF pellicle fabrication (100 mm × 100 mm with 21 nm thick) using camphor. During the transfer process, the elimination of Ni substrate and camphor results in a remarkable visual variation. Reproduced with permission. [25] Copyright 2022, The Authors, published by the Royal Society of Chemistry.
www.advmatinterfaces.de from 44.7 to 27.9 nm. The surface interaction of acetylene gas on Cu allowed the homogeneous growth of graphene to form the NGF. However, lots of defects are present on the basal plane of the graphene (enhancing the D-band intensity with increasing growth time). The NGF grown on both Ni foil and film exhibited a high degree of crystallinity (I D /I G ratio is around 0.02) (Figure 10b,c). It was found that the key parameter of uniform seed formation and lateral growth of graphene layers was carbon solubility/precipitation on Ni during heating and cooling, respectively. As discussed in the section on NGF growth on Ni foil (Figure 4b,c), the grain boundaries and step edges of Ni are the places where the carbon precipitation begins during cooling, resulting in the formation of thicker NGF around these Ni structures.

Optical, Thermal, and Chemical Properties of NGF
Reflecting EUV light from a photomask, namely a reticle, is mainly utilized for EUVL. As the EUV laser will be passed through the pellicle twice, the transmittance of the NGF pellicle at 13.5 nm (EUV) must be greater than 90%. The transmittance of freestanding NGF was measured using UV-vis-NIR spectroscopy (@550 nm, blue dots) and a homemade EUV transmittance measurement tool (@13.5 nm, red dots) as shown in Figure 11a. There are two fitting lines (red and blue dotted lines) that can be expressed to i) the equation for EUV [Equation (2)] where N · d is the thickness of the film, λ is the wavelength of the EUV, the optical constant for the EUV regime β ≈ (r 0 λ 2 /2π)n c f 2c (0) (where r 0 is classical electron radius, n c is the atomic density of NGF, and f 2c (0) is the imaginary part of the atomic scattering factor of carbon at a normal incident angle [38] ) and ii) the equation for visible light [Equation (3)] where κ = 1.3 is the extinction coefficient of graphite at λ = 550 nm, and t is the thickness of the NGF. It is noteworthy that in Figure 11a, we can extract the thickness and EUV transmittance of NGF using the simple measurement of transmittance at visible region (e.g., ≈51% of transmittance @550 nm → 18 nm of NGF and 92% of EUV transmittance). The thickness validation was performed by atomic force microscope (AFM).
Graphite, in general, has superior thermal properties (≈5000 W m K −1 ), [39] and can withstand high EUV absorption. Based on the computational simulation of the basal plane temperature increment of EUV pellicle materials such as graphene,  . Reproduced with permission. [16] Copyright 2017, Elsevier B.V.
www.advmatinterfaces.de silicon, and silicon nitride after 250 W of EUV laser exposure, NGF pellicle (323 °C) was found to have the lowest peak temperature rise after being exposed to 250 W of EUV laser, while silicon (≈700 °C) and silicon nitride (≈865 °C) were found to have the highest. [1] Using an 800 nm laser (infrared, IR), which has a larger absorption in the pellicle than the EUV laser, and delivers adjustable power density from (5-10 W cm −2 ), we demonstrated the thermal characteristics of the NGF pellicle. It is important that the NGF is freestanding on the frame during the measurement. The thermal image of the NGF pellicle with 5 W cm −2 of IR laser irradiation is depicted in Figure 11b. The center of the pellicle is illuminated with an IR laser, and ≈92% of the total intensity of the incident beam is focused within 2 mm of the center of the beam. Far from the center, the surface temperature of the NGF pellicle gradually decreases from 170 to 70 °C. As shown in Figure 11c, we also characterized the surface temperature and transmittance of NGF pellicle as laser power density (5-10 W cm −2 ) and laser exposure time (0-24 h) increased. The maximum temperature was 267 °C at the power density of 10 W cm −2 , which is still lower than in Si-based pellicles. The change in EUV transmittance after 24 h of laser irradiation was negligible. Similar to the temperature increase by EUV laser exposure condition shown in Figure 11c, we simulated the situation of delivering severe thermal energy (from 300 to 700 °C generated by rapid thermal annealing system) to NGF in EUV environment (3 Pa, H 2 ) without EUV laser (Figure 11d).
The chemical properties of the NGF pellicle have also been verified, as the NGF pellicle may suffer from the weak H 2 plasma generated by the EUV laser in the hydrogen environment in the chamber. NGF is degraded in two ways: i) by thermally reactive H 2 and ii) by H 2 plasma. The I D /I G ratio of each Raman spectrum remained nearly the same up to 500 °C as that of the as-grown NGF (0.02). At 700 °C, the ratio increased to 0.08 indicating that the reaction between hydrogen and graphene layer at the high temperature caused damage to the graphene layer. However, it is worth noting that NGF is only heated to ≈170 °C at the power density of 5 W cm −2 required for the EUVL, ensuring that there is no hydrogen damage on the thermally reactive NGF surface. We also characterized the NGF damage induced by H 2 plasma that is generated by the ionization of hydrogen-containing gas due to the absorption of high-energy EUV photons (92 eV). The H 2 plasma condition was maintained at 200 W and 10 Pa in the chamber, which is much harsher than in a real EUV chamber. [35] In Figure 11e, H 2 plasma damage to the NGF pellicle was confirmed by the increase of I D /I G ratio from 0.09 to 0.21 (130% increment of the defect), which can be reduced by deposition of a TiN capping layer (3 nm) on the NGF pellicle (0.23-0.26, 13% increment of the defect).

EUV Transmittance Mapping and Bulge Test of NGF
EUVL relies heavily on the uniformity of NGF pellicle over a large area (110 mm × 140 mm) because differential reflective-EUV light from a reticle, induced by thickness nonuniformity of NGF pellicle, permits hindering the fidelity of patterns on the wafer surface. A nonuniformity of 0.13% relative standard deviation of transmittance is desired, with a three-sigma value of less than 0.4%. The EUV transmittance of the 30 mm × 30 mm NGF pellicle has an average value of 87.2% and a Figure 10. Raman spectra of NGF synthesized on a) Cu foil, Reproduced with permission. [16] Copyright 2017 Elsevier B.V. b) Ni foil, Reproduced with permission. [2] Copyright 2015, The Royal Society of Chemistry. c) Ni film on Si substrate, Reproduced with permission. [24] Copyright 2021, IOP Publishing. www.advmatinterfaces.de relative standard deviation of 0.78%, as shown In Figure 12a. Based on the transmittance for EUV in Figure 11a, the thickness of the NGF is calculated to be 21 nm. In general, 0.2% of EUV transmittance is considered as the absorption of monolayer graphene, so 0.78% can be interpreted to reflect the three or four graphene layers difference on the surface of NGF.
Finally, a pellicle with outstanding mechanical properties of more than 1 TPa is necessary. [40] Mechanically exfoliated graphene has Young's modulus of 1 TPa, [12] indicating that NGF is the best material for a pellicle. The bulge test method is representative of measuring the mechanical properties of a broad area thin film including NGF pellicle. For example, the bulge test of nanometer-thick SiN x films enables confirmation that the results are comparable to those of a traditional tensile test. [41] The measurement tool for the bulge test of the NGF pellicle, [42] has an inner diameter and height of 200 mm and 100 mm, respectively, as shown in Figure 12b. A confocal displacement sensor (ZW-S5010) was used to measure the sample deflection. The NGF pellicle was prepared on the circular frame, and Young's modulus of NGF was measured to be 106 GPa, which is larger than that of bulk graphite. [43]

Perspective
Among practical pellicle materials, including carbon nanomaterials, Si, SiN x , Si-Mo-Nb multilayers, etc., NGF is considered as the next generation of EUV pellicle. To realize a full-size NGF pellicle for EUVL, we overcame substantial obstacles, including large-scale NGF growth and transfer method without sagging NGF on the frame. However, the nonuniformity of NGF (our best: 3σ = 2.34% at 100 mm × 100 mm of size Figure 11. Optical, thermal, and chemical properties of NGF pellicle. a) Transmittance of freestanding NGF pellicle at 13.5 and 550 nm, b) thermal properties (temperature of NGF exposed to laser) of NGF, c) NGF temperature with a laser power density (5 to 10 W cm −2 ) and transmittance of NGF after EUV exposure, d) Raman spectra of NGF after annealing in the hydrogen environment and I D /I G ratio, Reproduced with permission. [2] Copyright 2015, The Royal Society of Chemistry. e) Protection of NGF with a capping layer (TiN) against the H 2 plasma. Reproduced under the terms of the CC-BY license. [35] Copyright 2022, The Authors, published by IOP Publishing.

www.advmatinterfaces.de
(target: ≤ 0.4%)), strong mechanical properties of NGF (our best: 106 GPa (target: 1 TPa), and chemical tolerance of NGF against H 2 plasma (negligible additional defect formation) are still key challenges. Further breakthroughs in NGF synthesis, as well as NGF properties such as strong mechanical strength, chemical tolerance, uniformity, and defect-free, are required to realize the implementation/commercialization of NGF pellicle in the industry. In particular, i) the grains of Ni catalyst always affect the uniformity of NGF after CVD growth, necessitating precise strategies for forming a single crystal of Ni film with a specific lattice plane and controlling carbon precipitation by adding an intermediate cooling step for large NGF pellicle with a thickness of less than 20 nm. ii) Because the grain boundaries of NGF can be fracture-initiating defects under mechanical stress, the advanced CVD technique is required for growing single crystal NGF without point and line defects. iii) Finally, the chemical tolerance and mechanical strength can be enhanced by applying single crystal nanometer-thick capping layers with a good lattice match to NGF, such as SiC, B 4 C, TiN, etc.

Conclusion
The worldwide EUV market is expanding as a result of increased demand for IC chips in the electronics, automotive, military, and AI computing sectors. The demand for graphene-based pellicles (NGF pellicles), in particular, is soaring because such material may add new capabilities to IC chips by enhancing the mechanical, optical, and chemical properties. The central issue, however, is integrating such NGF into a large frame using a current or modified transfer process. We have reviewed only NGF pellicles in this study, concentrating on the scalability and their integration strategies into freestanding form. To fully realize the benefits of NGF pellicles, further study of growth, transfer or transfer-free methods, and capping layer (protection layer of NGF) must be developed. His current research is focused on the area of graphene/graphite synthesis/transfer for EUV pellicle, engineering quantum dots for next-generation displays and smart lighting systems, and smart textile.