Evaluation of H2 Plasma‐Induced Damage in Materials for EUV Lithography

Ultrafine semiconductor fabrication by lithography has undergone a significant transition from deep ultraviolet (DUV) to extreme ultraviolet (EUV) processes, which presents new challenges. Specifically, the damage caused to components utilized in an EUV system, such as multilayer mirrors, reticles, and pellicles within lithography equipment, owing to EUV‐induced H2 plasma, is a critical issue that directly affects the process yield and equipment lifespan. To address these issues, it is crucial to establish an environment similar to that of EUV‐induced plasma and develop a method to evaluate the resulting damage. Accordingly, an evaluation method is developed for assessing the material damage caused by hydrogen radicals and ions in inductively coupled H2 plasma. In these systems, the electron density ranged from 5 × 108 to 3.5 × 1010 cm−3, the electron temperature ranged from 1 to 4 eV, and the ion energy ranged from several to tens of eV; these conditions closely align with the environment of an EUV‐induced H2 plasma. The damage to Mo2C, a potential EUV pellicle material, is quantitatively analyzed by measuring the fraction of the pore area and examining the chemical characteristics after exposing the samples to various plasma conditions, including electron density, gas pressure, and exposure time.


Introduction
In the semiconductor industry, lithography technology was introduced in 1958 with the invention of integrated circuits (IC).The earliest photolithography processes were performed using visible (436 nm) or ultraviolet (365 nm) light emitted from DOI: 10.1002/admi.202300867mercury lamps.In the 1990s, as IC feature sizes decreased to submicrometers, lithography with finer line widths was required, which was difficult to achieve with visible and ultraviolet light.3][4][5] With the introduction of DUV and advances in lithography techniques (e.g., ArF immersion), the IC feature size has continued to decrease until recently. [6,7]he most advanced semiconductor processes demand sub-nanometer nodes in the nanoelectronics industry, and the previous generation of 193 nm ArFimmersion lithography processes cannot satisfy this requirement. [7,8][11][12] With the emergence of EUV, the development of new EUV materials has been actively pursued.However, it faces new challenges owing to a very different lithography environment than before, such as material damage caused by photon-induced plasma. [13]As EUV light is emitted from the source in the EUV lithography system, high-energy photons are absorbed by the background gas, and the absorbed photons can cause photoionization to create EUV-induced plasma, as shown in Figure 1. [14]Unlike the previously generated ArF 193 nm exposure equipment that used argon as a background gas, the inside of the EUV equipment with a wavelength of 13.5 nm is a H 2 plasma environment because H 2 gas with a pressure of 1-200 mTorr is used as the background gas.[26] Therefore, for a stable and reliable EUV lithography process, a method for the quantitative evaluation of tools and materials in the lithography equipment using H 2 plasma is required.However, because EUV sources are very expensive and high power is required to generate EUV-induced H 2 plasma in actual lithography processes, studies on the quantitative evaluation of the material damage caused by EUV-induced H 2 plasma are strongly limited.
An alternative is to use another type of plasma source with a relatively low setup cost and high-power efficiency to create a plasma environment similar to that inside lithographic equipment.][29] However, the ion energy incident on the electrode in a CCP is several hundred eV because of its high sheath potential, which is very high compared to the ion energy of several eV in EUV-induced H 2 plasma, and it is difficult to independently control the ion energy and electron density. [30]On the other hand, unlike the CCP, the inductively-coupled plasma (ICP) is a method of generating plasma by heating electrons with an electric field induced inside the chamber by the RF current flowing through the antenna. [31]Because ICP has a high-power efficiency, it is possible to obtain an electron density of 10 8 to 10 10 cm −3 , and an electron temperature of several eV can be achieved with a relatively low power compared to that achieved using EUV sources. [32,33][36] Therefore, an ICP source is a candidate system that can inexpensively and effectively generate the H 2 plasma environment similar to that of the lithography process.
In this study, we developed an efficient material damage evaluation system for H 2 plasma using ICP, and observed the damage to the material that could be applied to the EUV process or equipment.The plasma parameters of the developed system were measured in using a wave-cutoff probe, [37] Langmuir probe, and retarding field energy analyzer.To observe the effect of the H 2 plasma environment on the lithography tool, a Mo 2 C sample, which can be used as a barrier layer for multilayer mirrors or pellicle materials, [38] was exposed to plasma for various processing times, electron densities, and pressures.The damage to the samples due to plasma exposure was observed using fieldemission scanning electron microscopy (FE-SEM), and the compositional changes were analyzed by X-ray photoelectron spectroscopy (XPS).

Plasma Parameter Measurement
To verify that the plasma in the developed system is similar to the EUV-induced H 2 plasma environment, the electron density, effective electron temperature, electron energy probability function (EEPF), ion energy, and hydrogen radical density were measured under various input power and gas pressure conditions.

Electron Density
The absolute electron density measured using the wave-cutoff method is shown in Figure 2. Two antennas of the cutoff probe were inserted 72 mm above the stage and positioned at the center of the chamber to measure the plasma bulk.At a pressure of 10 mTorr (black square), as the input power increases from 50 to 800 W, the electron density increases from 4.6 × 10 8 to 1.6 × 10 10 cm −3 .From the energy balance equation, the electron density n 0 = P abs ∕eu B A eff  T , where P abs , u B , A eff and  T is the total absorbed power, Bohm velocity, effective loss area, and total energy lost per electron-ion pair lost, respectively.Because the electron density is proportional to the absorbed power P abs , it is expected that the electron density increases as the input power increases, and the measurement results show a corresponding trend.For the input power of 50 to 800 W, the electron density is 5.8 × 10 8 to 3.9 × 10 10 cm −3 at 20 mTorr (red circle), which is overall higher than at 10 mTorr, which can be interpreted as a result of the effective loss area decreasing with increasing pressure.As the pressure further increased and the mean free path of the electrons decreased, most of the ionization due to electron collision occurred near the antenna, and the diffusion of plasma decreased.Consequently, even if the electron density near the antenna increases, the electron density at a position far from the antenna may decrease.Accordingly, the electron density at a pressure of 100 mTorr (blue triangle) is 7.4 × 10 8 to 3.6 × 10 10 cm −3 , which is slightly lower than that at 20 mTorr.The electron density in EUV-induced plasma is ≈10 8 to 5 × 10 9 cm −3 , [27,39] and the electron density in developed system is 5 × 10 8 to 5 × 10 9 cm −3 at ICP power of 250 W or less, which is a good match.Moreover, because higher-density plasma can be generated by increasing the ICP power, it can be applied in high-power EUV environments corresponding to next-generation EUV equipment.

EEPF and Effective Electron Temperature
Figures 3 and 4 show the EEPF and effective electron temperature, respectively, measured using a Langmuir probe.The location of the Langmuir probe was the same as that of the cutoff probe.As shown in Figure 3a, for an input power of 50-800 W at 10 mTorr, the EEPF exhibits a Maxwellian distribution, indicating that the energy exchange between electrons is sufficient to reach thermal equilibrium.The effective electron temperature obtained by integrating the EEPFs was 3.5 eV as shown in Figure 4. Figure 3b shows the EEPF measured at 20 mTorr.Figure 3b shows that the EEPF in the developed system evolves from Maxwellian to bi-Maxwellian-like as the pressure increases to 20 mTorr, which can be explained by the effect of various inelastic collisions in the molecular gas.In H 2 plasma, a low-energy electrons are mainly generated by two mechanisms: low-energy electrons generated by the ionization process by electron-neutral collisions, and the energy of energetic electrons dissipated due to inelastic collisions.As the pressure increased, the electron density increased because the number of electron-neutral collisions increased.In addition, the energy loss of high-energy electrons increases in H 2 plasma because it has a number of inelastic reactions, such as rotational and vibrational excitations, electronic excitations, and dissociation of hydrogen molecules. [40]Considering the point at which the slope of the EEPF changes, vibrational and rotational excitations with large cross-sections near 3 eV [40] play major roles in the rapid energy loss of energetic electrons, which is consistent with the results of a previous study. [41]The effective electron temperature decreases to ≈2 eV owing to the increase in the population of low-energy electrons in the EEPF with increasing pressure, as shown in Figure 3b and 4. As the pressure was further increased to 100 mTorr at a power of 250 W, the EEPF evolved into a non-Maxwellian distribution with a deep hole near 3 eV, as shown in Figure 3c.The electron temperature decreases to ≈1 eV, and the electron density also decreases because the high-energy tail rapidly decreases in the process of the EEPF evolving into a strongly modified distribution as inelastic collisions increase further.As the power increases above 300 W, the deep hole changes into a plateau or puddle-like structure, which can be understood to be mainly due to the increase in electronelectron collisions. [42]The electron-electron collision frequency is proportional to the electron density and inversely proportional to the electron temperature. [43,44]As shown in Figure 4, the electron temperature at 100 mTorr is 1 eV, the lowest for all experimental pressures, and as shown in Figure 2, the electron density increases sharply above 300 W, resulting in an increase in electron-electron collisions.Consequently, the high-energy electrons decreased, and the low-energy electrons increased, resulting in a flattened EEPF profile near 3 eV.

Ion Energy
The ion energy distribution at the Mo 2 C sample location, as measured using the retarding field energy analyzer (RFEA-type) ion energy sensor, is shown in Figure 5.In the plasma, ions are generated in the bulk of the plasma and reach the sample or wall through the sheath.The ions are accelerated by the potential difference between the plasma and wall in the sheath region.In an ICP with a floating wall condition, the potential difference of the plasma and the floating wall is defined as |ΔV| ≈ T e ln( M 2m ) 1 2 , where T e is the electron temperature, M is the ion mass, and m is the electron mass. [30]Figure 5a shows the ion energy distribution with varying pressure in the H 2 plasma with an electron density of 10 10 cm −3 .The peak ion energy decreases from 11.5 to 6.1 eV as the pressure increases from 10 to 40 mTorr, as shown in Figure 5a, because the electron temperature decreases with increasing pressure (Figure 4). Figure 5b shows the ion energy distributions at various electron densities and a pressure of 20 mTorr.As shown in Figures 2 and 4, the electron density increased with increasing power at 20 mTorr, whereas the electron temperature increased slightly.Therefore, as the electron density increases from 5 × 10 9 to 3.5 × 10 10 cm −3 , the ion energy increases from 8.2 to 10.2 eV.At pressures above 50 mTorr, the ion flux reaching the sensor was low owing to the increase in particle collisions in the plasma, and reliable data could not be obtained.However, based on the measurement results, it can be expected that the ion energy is in the range of several tens of eV under the experimental conditions of this study, which is similar to the ion energy in EUV-induced H 2 plasma.

Hydrogen Radical
Figure 6 shows the H and H 2 densities measured by a quadrupole mass spectrometer in the H 2 plasma.The reaction rate of H atom generation by electron-neutral collisions is a function of the plasma parameters as R H = n e n g k(T e ,), where n e is the electron density, n g is the neutral gas density, k is the dissociation rate coefficient, T e is the electron temperature, and  is the collision cross-section of the dissociation. [30]As shown in Figure 6a, in H 2 plasma with an electron density of 10 10 cm −3 , the gas density of the hydrogen molecules increased as the pressure increased from 10 to 100 mTorr, and the hydrogen radical density also increased.The number of hydrogen molecules increased with the chamber pressure, and the number of hydrogen atoms increased with the parent gas density.However, the degree of dissociation is obtained as n H ∕(n H + 2n H 2 ), and decreased from 12.0 to 5.44% with increasing pressure.The decrease in the dissociation rate with increasing pressure resulted from a decrease in the dissociation rate coefficient as the electron temperature decreased (see Figure 3).[47] As the electron density increases from 10 9 to 3.5 × 10 10 cm −3 at a pressure of 20 mTorr, the H density increases, and the H 2 density tends to decrease as shown in Figure 6b.This is because the electron temperature is constant, and under fixed electron temperature and gas pressure conditions, the reaction rate is proportional to the electron density.Accordingly, the degree of dissociation increases from 6.30 to 8.28%.
From these results, it was confirmed that H 2 plasma with an electron density of 5.8 × 10 8 to 3.9 × 10 10 cm −3 , an electron temperature of 1 to 3.5 eV, and an ion energy of several to tens of eV was generated, which is in good agreement with the lithography process environment.In addition, because the QMS measurement results correspond well with those of previous studies and theoretical predictions, it is predicted that the hydrogen radicals generated by EUV lithography equipment will be similar to those in our system.

Surface Damage Analysis
To observe the damage to the Mo 2 C samples due to the H 2 plasma, nanostructures of the samples were measured by the FE-SEM after exposing the H 2 plasma.and cross-sectional FE-SEM images of the reference sample.The sample consisted of a 300 nm thick SiO 2 on a Si wafer and a 100 nm thick Mo 2 C layer deposited thereon.The Mo 2 C layer has a thickness difference of up to ≈20 nm, depending on the location, and pores with diameters of ≈100 nm that reach the SiO 2 surface through the Mo 2 C layer are interspersed.Under all the experimental conditions, the thickness of the Mo 2 C layer did not change significantly.To observe the damage of the sample under various H 2 plasma conditions, the area fraction of the pores was calculated from the FE-SEM image.The pore area fractions of the samples exposed to plasma with an electron density of 10 10 cm −3 at pressures of 10, 20, and 100 mTorr for 20 min are shown in Figure 7c.At the pressures of 20 and 100 mTorr, the fraction of pore area had no discernable change, but it increased to 2.455% at 10 mTorr.This result can be interpreted by that is because the ion density remains almost equal to the electron density due to the quasi-neutrality of the plasma, whereas the ion energy increases to 11.5 eV as the pressure is decreased, resulting in increased damage to the sample surface by ion bombardment.The measurement results in the previous section, in which the hydrogen radical density decreased with decreasing pressure, support the interpretation that the main cause of surface damage is ion bombardment rather than the chemical reaction of the hydro-gen radicals.The pore area fraction and surface FE-SEM images of samples exposed to plasma with the electron densities of 10 9 , 10 10 , and 3.5 × 10 10 cm −3 for 20 min under the gas pressure of 20 mTorr are shown in Figure 7d.The pore area fraction hardly changed at the electron densities of 10 9 and 10 10 cm −3 , but the Mo 2 C was etched, and the pore area fraction increased by ≈2% compared to the reference sample at an electron density of 3.5 × 10 10 cm −3 .The hydrogen ion density increases with increasing electron density because of the quasi-neutrality of the plasma.This increase in the hydrogen ion flux can cause an increase in the pore area by ion bombardment.The electron temperature and ion bombardment energy measurements with increasing electron density introduced in the previous section support this interpretation.As the plasma exposure time of the Mo 2 C sample increased, the damage to Mo 2 C was further aggravated owing to the increase in ion bombardment time, and Figure 7e shows the corresponding result.In other words, the surface damage of the Mo 2 C samples under environmental conditions similar to those of EUV-induced H 2 plasma is mainly affected by the electron density and ion energy, which are related to the ion flux and exposure time.
XPS analysis was performed to observe chemical changes induced by the hydrogen plasma on Mo 2 C sample surfaces.The  Mo 3d spectrum of the reference sample is shown in Figure 8a.[50][51] After the reference Mo 2 C sample was exposed to plasmas with an electron density of 3.5 × 10 10 cm −3 at a pressure of 20 mTorr for 120 and 240 min, we also obtained Mo 3d spectra, as shown in Figure 8b,c, respectively.As the plasma exposure time increases, the Mo 3d spectra exhibit a decrease in the intensities of Mo 2+ state while two new doublet peaks are observed at 232.74 and 235.88 eV, and at 229.49 and 232.64 eV, corresponding to the Mo 6+ and Mo 4+ states, respectively.The peak position are in good agreement with the Mo 6+ and Mo 4+ states of fully oxidized MoO 3 [52,53] and MoO 2 , [54] respectively.Therefore, it suggests that under these plasma exposure conditions, part of the Mo 2 C sample surface is transformed into MoO 2 and MoO 3 , and the degree of transformation increased with the prolonged

Conclusion
An ICP-based material damage evaluation system was developed.In these system, the electron density of the plasma was in the range of 5 × 10 8 -3.5 × 10 10 cm −3 depending on the input power; the electron temperature was 1-4 eV depending on the pressure; the peak ion energy was 6.1 to 11.5 eV, which are very similar to that in the EUV-induced H 2 plasma environment.As an example of material damage evaluation, one of the candidates for the EUV pellicle, Mo 2 C, was exposed to H 2 plasma, and the damage to the samples was observed using FE-SEM and XPS.The samples were slightly etched under a low pressure and high electron density, which are conditions of a relatively high ion energy; the longer the exposure time, the more severe the compositional deformation.These results confirmed that the developed system can be effectively used to evaluation H 2 plasma induced damage to EUV lithography tools and materials, and therefore, this method can be used for quantitative evaluation of materials for EUV lithography.

Experimental Section
Damage Evaluation System: A schematic of the H 2 plasma system is shown in Figure S1 (Supporting Information).Up to 800 W of 13.56 MHz RF power was delivered from the power supply to the one-turn coil an-tenna through an auto-matching unit.H 2 gas (99.999% purity) was injected into the chamber and the flow rate was controlled at 20 sccm using a mass flow controller.The experimental pressures were 10, 20, and 100 mTorr, which were measured using a capacitive manometer gauge calibrated with a standard manometer at the Korea Research Institute of Standards and Science (KRISS).The pumping system consisted of a rotary vane pump and a turbo molecular pump, and the pressure was controlled using a throttle valve.The electron density was measured using the wave-cutoff method [37] using a cutoff probe connected to a network analyzer (8753ET, Agilent) with two coaxial cables.EEPF were measured using a single Langmuir probe with a reference ring.The ion energy was measured using a retarding-field energy analyzer (Semion RFEA system, Impedans), and the ion energy sensor was placed on the sample stage 200 mm below the top of the vacuum chamber.While the sample was exposed to the plasma, the cutoff probe was positioned 100 mm radially outward from the center to avoid shadowing effects and minimize plasma perturbation.A quadrupole mass spectrometer (QMS, Hiden PSM) was mounted on the chamber sidewall and evacuated using a differential pumping system consisting of a turbomolecular pump and a rotary pump.The QMS was maintained in vacuum at the 1 × 10 −7 Torr, as measured using a cold cathode gauge.An orifice with a diameter of 30 μm was placed between the process chamber and the QMS.During the experiments, the electron energy of the ionizer and emission current were set to 20 eV and 50 μA, respectively.The sample used for damage evaluation was fabricated through RT-CVD and consisted of a 300 nm SiO 2 layer on a Si wafer and a 100 nm Mo 2 C layer on it.The sample was loaded into the center of a cylindrical stage with a diameter of 250 mm using a load-lock chamber, and placed 200 mm away from the dielectric window.The surface nanostructure of the Mo 2 C sample after plasma exposure was measured by FE-SEM (S-4800, Hitachi) at 10 keV.The XPS analysis (VersaProbe II, Ulvac-PHI) was performed using an Al-K X-ray source with an anode voltage of 15 kV.The base pressure of the analysis chamber was 2 × 10 −7 mTorr, and the X-ray beam emission angle was 45°.

Figure 2 .
Figure 2. Electron density measured by the wave-cutoff method at pressures of 10, 20, and 100 mTorr in the H 2 plasma.

Figure 3 .
Figure 3. EEPF measured by the Langmuir probe at pressures of a) 10 mTorr, b) 20 mTorr, and c) 100 mTorr in the H 2 plasma.

Figure 4 .
Figure 4. Effective electron temperature at pressures of 10, 20, and 100 mTorr in the H 2 plasma.
Figure 7a,b show surface

Figure 5 .
Figure 5. Ion energy distribution measured by RFEA in the H 2 plasma a) with an electron density of 10 10 cm −3 at 10 to 40 mTorr, b) with an electron density of 5 × 10 9 to 3.5 × 10 10 cm −3 at 20 mTorr.

Figure 6 .
Figure 6.Hydrogen radical and molecular density measured by QMS in the H 2 plasma a) with an electron density of 10 10 cm −3 at 10 to 100 mTorr, b) with an electron density of 10 9 to 3.5 × 10 10 cm −3 at 20 mTorr.

Figure 7 .
Figure 7. a) Top-view and b) cross-sectional FE-SEM images of the reference sample.c) Top-view FE-SEM images and pore area fractions of the samples exposed to the plasma with different pressures, d) electron densities, and e) exposure times.

Figure 8 .
Figure 8. Mo 3d spectra of the a) the reference sample and samples exposed for b) 120 min and c) 240 min.