Method for preparing high-resolution EUV optical element by using electron beam lithography

Abstract

The invention belongs to the technical field of micro-nano machining, and particularly relates to a method for preparing a high-resolution EUV optical element through electron beam lithography. The method comprises the following steps: etching an optical element by using an electron beam lithography system, and introducing a region proximity effect correction strategy; the method comprises the following specific steps: preparing a silicon nitride film as a substrate, and spin-coating HSQ photoresist on the silicon nitride film; carrying out geometric region division on the to-be-exposed optical component graphic layout; at least dividing a central region, an edge region and a corner region; utilizing Monte Carlo simulation software to establish a function relationship between the graphic position and the required exposure dose so as to determine an independent exposure reference dose of each geometric region; an independent and compensated exposure reference dose is set for each geometric region based on dose deviation data obtained by simulation. According to the method, the pattern uniformity can be remarkably improved, and reliable preparation of a large-area and large-aspect-ratio nano structure is realized.

Description

Technical Field

[0001] This invention belongs to the field of micro-nano fabrication technology, specifically relating to a method for fabricating high-resolution EUV optical elements using electron beam lithography. Background Technology

[0002] As semiconductor manufacturing processes advance to smaller nodes, extreme ultraviolet (EUV) lithography has become a core driving force for industry development. Meanwhile, in cutting-edge research fields such as materials science, life sciences, and environmental science, EUV / soft X-ray microscopy based on synchrotron radiation sources is playing an increasingly important role. In these applications, high-performance optical components are crucial in determining system performance. Because the refractive index of materials in the EUV / soft X-ray band is close to 1, traditional refractive optical components fail; therefore, optical components in this band are mainly reflective or diffractive. Among them, diffractive optical components, such as Fresnel zone plates and diffraction gratings, have attracted much attention due to their ability to achieve high-resolution focusing and imaging.

[0003] To achieve higher optical efficiency, diffraction element materials are shifting from traditional metals (such as gold and iridium) to dielectric materials with lower absorption and stronger phase modulation capabilities. Hydrosilylsilsesquioxane (HSQ) is a high-performance inorganic negative electron photoresist that can be transformed into a stable dielectric material similar to silicon dioxide after electron beam exposure. It also possesses the potential for ultra-high resolution down to sub-10 nm, making it an ideal choice for fabricating EUV-band dielectric optical elements.

[0004] Electron beam lithography is currently the primary fabrication method capable of meeting the nanometer-level resolution requirements of EUV optical components. However, electron beam lithography faces an inherent physical challenge in practical applications—the proximity effect. When an electron beam bombards the photoresist, electrons undergo forward and backscattering, causing the exposed area to extend beyond the electron beam spot, affecting the accuracy of the pattern. To address this microscopic problem, the industry has developed mature proximity effect correction techniques. By segmenting the pattern and fine-tuning the exposure dose in different regions, dose variations caused by proximity features can be effectively compensated.

[0005] However, traditional proximity correction techniques fail when the desired pattern simultaneously meets the three conditions of large area (e.g., 1 mm × 1 mm), high density (e.g., half-cycle of 15 nm), and large aspect ratio (e.g., greater than 10:1). The fundamental reason is that traditional proximity correction techniques primarily address microscopic proximity effects caused by local, adjacent patterns, neglecting the macroscopic proximity effects caused by backscattered electrons accumulated over the entire large-area pattern. Monte Carlo simulations and experimental results both show that in a large-area periodic pattern, the cumulative energy density absorbed in the central region is significantly higher than that in the edge and corner regions. This macroscopic dose inhomogeneity results in overexposure in the central region and underexposure in the edge and corner regions, even when using proximity correction techniques in commercial software. This underexposure, especially in high aspect ratio structures, easily leads to line adhesion and collapse during subsequent development, ultimately causing the failure of the entire device.

[0006] Therefore, overcoming the macroscopic proximity effect in the processing of large-area, high-density, and high aspect ratio patterns, and achieving uniform and reliable replication of nanostructures throughout the entire pattern area, is a technical bottleneck that urgently needs to be solved in the current EUV optical element manufacturing field. Summary of the Invention

[0007] The purpose of this invention is to provide a method for fabricating high-resolution EUV optical elements using electron beam lithography, in order to solve the problems of reduced pattern quality and low yield caused by macroscopic dose unevenness when using existing electron beam lithography technology to fabricate large-area, high-density, high aspect ratio micro-nano patterns.

[0008] The present invention provides a method for fabricating high-resolution EUV optical elements using electron beam lithography. The method utilizes an electron beam lithography system to etch the optical elements, during which a region proximity effect correction strategy is introduced. The specific steps are as follows:

[0009] (a) Prepare ultrathin silicon nitride films with a thickness of 50-100 nm and an area of ​​1 mm × 1 mm to 4 mm × 4 mm;

[0010] (ii) Using a silicon nitride thin film as a substrate, a layer of HSQ photoresist is spin-coated on it;

[0011] (iii) Divide the pattern of the optical component to be exposed into geometric regions. Specifically, based on the macroscopic geometric position of the pattern, divide it into at least three regions, typically including a central region, an edge region, and a corner region where the dose difference is most significant;

[0012] (iv) Using Monte Carlo simulation software (GenISys' TRACER / BEAMER), a functional relationship between the graphic position and the required exposure dose was established to determine the independent exposure reference dose for each geometric region; the Monte Carlo simulation software was used to accurately calculate the three-dimensional absorbed energy density distribution of the entire pattern during electron beam exposure; through the simulation model, the dose deviation caused by the macroscopic proximity effect in different geometric regions was quantified. Specifically:

[0013] (1) Establish a simulation model, including:

[0014] (a) Substrate: SiN with a thickness of 50–100 nm X ;

[0015] (b) Photoresist: 100–400 nm HSQ;

[0016] (c) Accelerating voltage: 100 keV;

[0017] (d) Electron beam radius: 7 nm;

[0018] (e) Layout size: 100 μm × 100 μm to 1 mm × 1 mm;

[0019] (2) Obtain the average energy deposition at different locations, including:

[0020] Monte Carlo simulation output energy distribution diagram ( Figure 3 (Corresponding), and statistics are compiled separately, see Table 1.

[0021] Table 1. Output energy distribution of Monte Carlo simulation

[0022] ;

[0023] (3) Quantifying macroscopic dose deviation:

[0024] The dose deviation coefficient is defined as:

[0025] ; Where E_center is the average absorbed energy density of the central region. This represents the average absorbed energy density in the edge or corner region.

[0026] For example, taking the median value of E_edge (0.90) and the median value of E_corner (0.78) in Table 1, we can obtain: edge region: k_edge ≈ 1 / 0.90 ≈ 1.11; corner region: k_corner ≈ 1 / 0.78 ≈ 1.28.

[0027] (4) Calculate the regional compensation dose:

[0028] Let Dcenter_opt be the optimal exposure reference dose for the central region, then the target exposure reference dose for the i-th region is:

[0029] ;

[0030] Table 2 Monte Carlo Simulation Output Exposure Dose

[0031] ;

[0032] If the dose in the central region is D_center = 4000 μC / cm², then:

[0033] .

[0034] (v) Based on the dose deviation data obtained in step (iv), an independent, compensated exposure reference dose is set for each divided geometric region; relative to the reference dose of the central region, the reference dose of the edge region and the corner region is increased accordingly to make up for their insufficient energy absorption.

[0035] The following method was used to determine the exposure reference dose for each region:

[0036] (1) Determine the critical imaging dose D_th (e.g., 3000 μC / cm²) of the central region by independent dose matrix test.

[0037] (2) Calculate the optimal exposure dose D_center_opt according to the target linewidth (15–50 nm) of the design. Example: Target 28 nm gate line → D_center_opt = 4000 μC / cm².

[0038] (3) Calculate the compensation dose:

[0039] ; in, The deviation coefficient is obtained in step (iv).

[0040] (4) Generate a regional dose table.

[0041] (vi) In the electron beam lithography system, a zoned exposure mode is adopted, and a corresponding independent exposure reference dose is applied to each geometric region;

[0042] (vii) After exposure, develop in TMAH solution at 45-55 ℃ (preferably 50 ℃) (e.g., 3 min) and fix in deionized water (1 min).

[0043] In step (iv) of this invention, Monte Carlo simulation software is specifically used to accurately calculate the three-dimensional absorption energy density distribution of the entire pattern during electron beam exposure; through simulation, the dose deviation caused by macroscopic proximity effect in different geometric regions is quantified.

[0044] In this invention, the geometric region includes a corner region, and the exposure reference dose of the corner region is different from the exposure reference dose of the center region and the edge region.

[0045] In this invention, the height of the HSQ structure is 100 nm to 400 nm, the minimum line width is 15 nm (period of 30 nm) to 50 nm (period of 100 nm), and the aspect ratio is greater than 7:1.

[0046] In this invention, the area of ​​the optical component layout is from 100 μm × 100 μm to 1 mm × 1 mm.

[0047] In this invention, the development is performed using a tetramethylammonium hydroxide (TMAH) solution at a temperature of 40°C to 60°C.

[0048] In this invention, the optical components typically include Fresnel zone plates or diffraction gratings.

[0049] In this invention, the period of the diffraction grating is 56 nm and the line width is 15 nm - 100 nm.

[0050] In this invention, the outermost ring width of the Fresnel zone plate is 15 nm - 100 nm.

[0051] In this invention, the area of ​​the optical pattern is equal to or less than the effective area of ​​the silicon nitride thin film window on the substrate.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] (1) Significantly improves image uniformity: The regional proximity effect correction method of the present invention fundamentally solves the problem of dose unevenness caused by macroscopic proximity effect. By performing differentiated dose compensation on different regions, it ensures that the entire large-area image from the center to the edge can obtain uniform and optimal exposure, thereby completely eliminating the image collapse phenomenon in the edge and corner areas and ensuring the integrity and fidelity of the image.

[0054] (2) Reliable fabrication of large-area, high aspect ratio nanostructures: This invention enables the stable fabrication of dense nanostructures with feature sizes less than 15 nm and aspect ratios exceeding 10:1 on areas of 1 mm × 1 mm or even larger. This removes a key process obstacle for manufacturing EUV diffraction optical elements with larger numerical apertures and higher optical performance.

[0055] (3) High process reliability and wide process window: Through precise simulation and dose compensation, this method reduces the process sensitivity to exposure dose and widens the effective process window. Experiments show that the process is reliable in the range of 2000-4000 μC / cm². 2 High-quality graphs can be obtained over a wide dose range, demonstrating the high stability and reproducibility of the method.

[0056] (4) Improve the performance of the final optical components: Since the fabricated optical components have uniform patterns and complete structures, their key optical performance indicators such as diffraction efficiency and resolution are effectively guaranteed, which can meet the needs of high-end applications such as EUV mask defect detection and EUV photoresist performance evaluation. Attached Figure Description

[0057] Figure 1 This is a schematic diagram of the preparation method of the present invention.

[0058] Figure 2 The images show physical models of the ultrathin silicon nitride membranes prepared according to this invention and the transmittance of silicon nitride in the EUV band. The thickness is 50 nm, and the areas are 1 mm × 1 mm, 2 mm × 2 mm, and 4 mm × 4 mm, respectively.

[0059] Figure 3 This is a Monte Carlo simulation diagram of the uneven absorption energy density distribution in a large-area graphic before the use of the regional proximity effect correction method.

[0060] Figure 4 This is a schematic diagram of the region division for the regional proximity effect correction method of the present invention, showing the center, edge and corner regions.

[0061] Figure 5 This is a diagram showing the electron energy distribution of the outer ring of a zone plate after the region proximity effect correction method of the present invention.

[0062] Figure 6 SEM images of gratings prepared without using the region proximity effect correction method.

[0063] Figure 7 The SEM image (200 μm × 200 μm) of the grating was prepared using the regional proximity effect repair method of the present invention.

[0064] Figure 8The SEM image (400 μm × 600 μm) of the grating was prepared using the regional proximity effect repair method of the present invention.

[0065] Figure 9 This is a SEM image of the outermost ring region of a Fresnel zone plate with a resolution of 50 nm and a diameter of 1 mm, successfully prepared using the method of this invention. Detailed Implementation

[0066] The present invention will be further explained below with reference to the accompanying drawings and embodiments.

[0067] The core idea of ​​this invention lies in recognizing the challenges of electron beam lithography in fabricating large-area, high-density patterns, and proposing a novel method for correcting regional proximity effects based on macroscopic geometric positions. Traditional proximity effect correction methods primarily address the "microscopic" proximity effect caused by backscattered electrons with radii within a few micrometers, achieved by adjusting the dose of a single pattern or a small cluster of patterns. However, for a large pattern larger than 100 μm × 100 μm, especially over 500 μm, its center is affected by the cumulative effect of backscattered electrons contributed by all exposure points throughout the pattern region (up to 100 µm away), while the cumulative effect on the edge positions is much smaller. This "macroscopic" cumulative effect is the fundamental reason for the "high center, low edge" dose distribution across the entire pattern, and it is something that traditional proximity effect correction methods cannot solve.

[0068] The region proximity effect correction method of this invention is precisely designed to solve this macroscopic problem. It first uses Monte Carlo simulation, a powerful quantitative analysis tool, to accurately predict the actual absorbed energy at different locations (center, edge, and corner areas) within a large-area pattern under a specific substrate and photoresist system. Thus, after actual exposure, the effective energy absorbed by all regions will tend to be consistent, thereby achieving a uniform pattern.

[0069] Example 1: Fabrication of a large-area, high-aspect-ratio diffraction grating

[0070] This embodiment fabricates a diffraction grating for EUV interference lithography, with the following technical specifications: area 400 μm × 600 μm, period 56 nm (i.e., 28 nm line / 28 nm gap), and height 350 nm (aspect ratio approximately 12.5:1). The process flow is as follows: Figure 1 As shown, the detailed preparation steps are as follows:

[0071] (1) Substrate preparation: Select a substrate with a thickness of 50-100 nm of silicon nitride (SiN). X The silicon wafer with the thin-film window serves as the substrate, such as Figure 2 As shown.

[0072] (2) Coating: Spin coat a layer of HSQ photoresist with a thickness of 350 nm onto the silicon nitride film, and then bake it in an oven at 180°C for 1 hour.

[0073] (3) Graphic design and regional proximity effect correction settings:

[0074] a) Design a grating pattern with an area of ​​400 μm × 600 μm and a period of 56 nm. To prevent the lines from collapsing, reinforcing ribs can be designed along the direction perpendicular to the grating lines.

[0075] b) In electron beam lithography data preparation software (such as BEAMER), divide the large-area pattern into eight main geometric regions, including the center, four sides, and four corners, such as... Figure 4 As shown.

[0076] c) Run a Monte Carlo simulation to calculate the temperature at 4000 μC / cm. 2 The energy density differences in different regions at the baseline dose. Based on the simulation results, regional compensation doses were set as follows: 100% (base dose) for the central region, 105%-120% for the edge region, and 110%-130% for the corner region. Specific values ​​were fine-tuned based on the simulation results. Figure 4 As shown.

[0077] d) To achieve the final linewidth of 28 nm, considering process variations, the designed exposure linewidth is set to 18 nm, such as... Figure 5 As shown.

[0078] (4) Electron beam lithography: Exposure is performed using a high-resolution electron beam lithography machine. The system will automatically apply different exposure doses to different regions based on the region proximity effect correction parameters set in step 3. The base dose is set to 4000 μC / cm. 2 .

[0079] (5) Development: Immerse the exposed sample in a 0.25 mol / L TMAH developer solution and develop it for 3 minutes in a 50°C water bath. Then rinse with deionized water for 1 minute and dry with nitrogen.

[0080] (6) Results Analysis: Scanning electron microscopy was used for detection. The results showed that the grating pattern within the entire 400 μm × 600 μm area was uniform, with clear lines, steep sidewalls, and no collapse or adhesion phenomena. Figure 7 and Figure 8 As shown. Results for the same sample not prepared using the method of this invention are as follows. Figure 6 As shown.

[0081] Example 2: Preparation of high-resolution Fresnel zone plates

[0082] This embodiment prepares a Fresnel zone plate for EUV mask defect detection, with the following technical specifications: diameter 1 mm, outermost ring width 50 nm, and height 200 nm. Detailed preparation steps are as follows:

[0083] (1) Substrate preparation: Select a substrate with a 50 nm thick silicon nitride (SiN) substrate. X The silicon wafer with the thin-film window serves as the substrate.

[0084] (2) Coating: Spin-coating a layer of HSQ with a thickness of 200 nm onto the silicon nitride film, and then soft-baking in an oven at 180°C for 1 hour.

[0085] (3) Graphic design and regional proximity effect correction settings:

[0086] a) Design a zone plate pattern with a diameter of 1 mm and an outermost ring width of 50 nm. To prevent the lines from collapsing, reinforcing ribs can be designed along the direction perpendicular to the grating lines.

[0087] b) In the electron beam lithography data preparation software, the large-area pattern is divided into 8 main geometric regions, including 8 annular regions from the outside in. c) A Monte Carlo simulation is run to calculate the energy density differences between the regions at a base dose of 4000 μC / cm². Based on the simulation results, the regional compensation doses are set as follows: 100% (base dose) from the center region to the outermost ring region, 110%, 120%, 130%, and 170% up to the outermost ring region. The specific values ​​are fine-tuned based on the simulation results.

[0088] (4) Electron beam lithography: Exposure is performed using a high-resolution electron beam lithography machine. The system will automatically apply different exposure doses to different regions according to the region proximity effect correction parameters set in step 3. The base dose is set to 4000 μC / cm2.

[0089] (5) Development: Immerse the exposed sample in a 0.25 mol / L TMAH developer solution and develop it for 3 minutes in a 50°C water bath. Then rinse with deionized water for 1 minute and dry with nitrogen.

[0090] (6) Results Analysis: Scanning electron microscopy (SEM) was used for inspection. The results showed that the zone plate pattern within a 1 mm diameter area had clear lines, steep sidewalls, and no collapse or adhesion. The measured line width was approximately 50 nm and the height was 200 nm, fully meeting the design requirements. Figure 9 As shown.

[0091] In summary, this invention, by introducing an innovative regional proximity effect correction method and combining it with the high resolution characteristics of HSQ photoresist and the stability of high-temperature development process, successfully overcomes the long-standing technical bottleneck in the fabrication of large-area, high-density, and high aspect ratio patterns in the field of micro-nano fabrication. This provides a reliable and efficient technical approach for the research and development and production of high-performance EUV optical core components.

Claims

1. A method for fabricating high-resolution EUV optical elements using electron beam lithography, characterized in that, Optical components are etched using an electron beam lithography system, during which a region proximity effect correction strategy is introduced; the specific steps are as follows: (a) Prepare ultrathin silicon nitride films with a thickness of 50-100 nm and an area of ​​1 mm × 1 mm to 4 mm × 4 mm; (ii) Using a silicon nitride thin film as a substrate, a layer of HSQ photoresist is spin-coated on it; (iii) Divide the pattern of the optical component to be exposed into geometric regions. Specifically, based on the macroscopic geometric position of the pattern, divide it into at least three regions, including the central region, the edge region, and the corner region where the dose difference is most significant. (iv) Using Monte Carlo simulation software, establish the functional relationship between the graphic position and the required exposure dose to determine the independent exposure reference dose for each geometric region; specifically, use Monte Carlo simulation software to accurately calculate the three-dimensional absorption energy density distribution of the entire pattern during electron beam exposure; through the simulation model, quantify the dose deviation caused by the macroscopic proximity effect in different geometric regions. (v) Based on the dose deviation data obtained in step (iv), an independent and compensated exposure reference dose is set for each divided geometric region; relative to the reference dose of the central region, the reference dose of the edge region and the corner region is increased accordingly to make up for their insufficient energy absorption. (vi) In the electron beam lithography system, a zoned exposure mode is adopted, and a corresponding independent exposure reference dose is applied to each geometric region; (vii) After exposure, develop in TMAH solution at 45-55 ℃ and fix in deionized water.

2. The method according to claim 1, characterized in that, The height of the HSQ structure described in step (ii) is 100 nm to 400 nm, the minimum line width is 15 nm to 50 nm, and the aspect ratio is greater than 7:

1.

3. The method according to claim 1, characterized in that, The area of ​​the optical component pattern in step (iii) is from 100 μm × 100 μm to 1 mm × 1 mm, and the area of ​​the optical component pattern is equal to or less than the effective opening area of ​​the silicon nitride thin film window on the substrate.

4. The method according to claim 1, characterized in that, The exposure reference dose in the corner region is increased by 10% to 35% compared to the center region and the edge region, respectively.

5. The method according to claim 1, characterized in that, Step (four) specifically includes: (1) Establish a simulation model, including: (a) Substrate: SiN with a thickness of 50–100 nm X ; (b) Photoresist: 100–400 nm HSQ; (c) Accelerating voltage: 100 keV; (d) Electron beam radius: 7 nm; (e) Layout size: 100 μm × 100 μm to 1 mm × 1 mm; (2) Obtain the average energy deposition at different locations, including: The energy distribution diagrams output by the Monte Carlo simulation are shown in Table 1, and the statistics are presented separately. Table 1. Output energy distribution diagram of Monte Carlo simulation ； (3) Quantifying macroscopic dose deviation: The dose deviation coefficient is defined as: ； Where E_center is the average absorbed energy density of the central region. This represents the average absorbed energy density in the edge or corner region. (4) Calculate the regional compensation dose: If the dose in the central region is D_center = 4000 μC / cm², then: ， 。 6. The method according to claim 1, characterized in that, In step (v), based on the energy deposition simulation data from step (iv), the exposure baseline dose for each region is determined using the following method: (1) Determine the critical imaging dose D_th of the central region based on the dose matrix experiment; (2) Determine the optimal exposure reference dose D_center_opt for the central region according to the design target linewidth (15–50 nm); (3) Calculate the target exposure reference dose for each region according to the relationship D_i=D_center_opt×k_i given in claim 5; (4) Based on the target exposure reference dose, and in combination with the stability of the exposure system and the development process, a dose process window of 5% to 10% is set for each area to form a regional dose table as a reference for setting during zoned exposure.

7. The method according to any one of claims 1-6, characterized in that, The optical component is a Fresnel zone plate or a diffraction grating.

8. The method according to claim 6, characterized in that, The diffraction grating has a period of 56 nm and a line width of 15 nm to 100 nm; the outermost ring of the Fresnel zone plate has a width of 15 nm to 100 nm.

Images