Mask blank glass substrate

Abstract

Providing a glass substrate for mask blanks that has a highly flat, low-defect, and low-roughness surface, sufficiently high reflectivity of EUV light after the deposition of a reflective film, and is suitable as a substrate for EUVL mask blanks for state-of-the-art applications. A glass substrate for mask blanks having a maximum value of an annular average power spectral density of 1000 nm or less, obtained by measuring the surface shape of a 10 µm × 10 µm region with an atomic force microscope and a spatial frequency of 0.1 µm⁻¹ or more and 20 µm⁻¹ or less.

Description

Technology Field

[0001] The present invention relates to a glass substrate for mask blanks, and more specifically, to a glass substrate for mask blanks used in state-of-the-art EUV lithography. Background Technology

[0002] In recent years, EUV lithography (hereinafter abbreviated as "EUVL"), an exposure technology using EUV (Extreme Ultra Violet) light, has been attracting attention in order to realize the formation of fine patterns exceeding those of conventional ultraviolet photolithography.

[0003] EUV light is light in the wavelength band of the soft X-ray region or vacuum ultraviolet region with a wavelength of about 0.2 to 100 nm, and reflective masks are considered promising as transfer masks used in EUVL.

[0004] The substrate used for this reflective mask requires a surface with significantly reduced surface roughness, flatness, and the number of microdefects.

[0005] In particular, increasing the reflectivity of EUV light from reflective masks is important for improving throughput in the EUVL exposure process, and to achieve this, it is necessary to reduce the surface roughness of the substrates for reflective mask blanks.

[0006] It is common practice to use an atomic force microscope to measure the surface roughness of a substrate. For example, Non-Patent Literature 1 reports that there is a correlation between the surface roughness (RMS) of a substrate measured by an atomic force microscope in a square area with sides of 10 μm and the reflectance of EUV light after the formation of a reflective film.

[0007] Meanwhile, in Patent Document 1, the surface roughness (RMS) obtained by measuring a 1㎛×1㎛ region with an atomic force microscope is 0.15 nm or less, and the spatial frequency is 1㎛ -1 10㎛ or more -1Power spectral density (PSD) in the range below 10 nm 4 A reflective mask substrate and a processing method thereof have been reported.

[0008] In addition, Patent Document 2 describes a spatial frequency of 1×10 in the range of 0.14×0.1 mm. -2 ㎛ -1 The above PSD is 4×10 6 nm 4 Less than or equal to, and also the spatial frequency of 1 μm in a square with a side length of 1 μm -1 A PSD in the range of 10 nm 4 A substrate with the following characteristics has been reported.

[0009] In addition, Patent Document 3 reports a glass substrate for mask blanks in which the removal of foreign substances present on the surface is facilitated by controlling the aspect ratio of the surface properties Str(s=0.2)(ISO025178-2) to 0.30 or higher. Prior art literature

[0010] International Publication No. 2013 / 146990, International Publication No. 2014 / 104276, Japanese Patent Publication No. 2016-143791

[0011] Noriaki Kamitaka, Katsuhiko Murakami, PF NEWS Vol.26, No.1, pp.24-25, 2008 The problem to be solved

[0012] However, since the reflectance of EUV light depends on the spatial period of surface roughness, just like the diffraction phenomenon of X-rays, it cannot be said that high reflectance is necessarily obtained after the formation of the reflective film by managing surface roughness by RMS alone, as in the description of Non-Patent Literature 1.

[0013] Furthermore, recent studies have shown that the factor significantly affecting reflectance is 0.1 μm, which has a smaller spatial frequency than reported in Patent Documents 1 and 2. -1It was found that if the PSD in this range is not precisely controlled, a stable reflectance cannot be obtained.

[0014] In addition, titanium-doped glass substrates, which are commonly used as substrates for EUV reflective mask blanks, may have a distribution of titanium concentration in the material, and due to this titanium concentration distribution, a periodic structure in the height direction may occur in a relatively macroscopic area of ​​the millimeter order on the surface of the substrate. This periodic structure has a height of several nm to about 10 nm, and in reflective mask blank substrates where flatness (TIR) ​​of less than 30 nm is required, it becomes a factor that degrades flatness. In this regard, defining only the PSD in the spatial frequency range described in Patent Documents 1 and 2 is insufficient to satisfy the various characteristics required for reflective mask blank substrates comprehensively, and thus a wider range of PSD control is required.

[0015] Meanwhile, in the technology of Patent Document 3, although it is certainly possible to remove foreign substances and reduce micro-defects, as described above, it is difficult to satisfy the necessary characteristics of reflectance or flatness unless the PSD is observed over a wider range of spatial frequencies and the surface structure is controlled comprehensively.

[0016] The present invention is made in consideration of the above circumstances and aims to provide a glass substrate for mask blanks that is highly flat, has a low defect, has a surface with low surface roughness, has a sufficiently high reflectivity of EUV light after the deposition of a reflective film, and is suitable as a substrate for EUVL mask blanks for state-of-the-art applications. means of solving the problem

[0017] The inventors, after careful consideration to achieve the above objective, discovered that a substrate in which the maximum value of the annular average power spectral density calculated by measuring a 10㎛ × 10㎛ area within the surface by an atomic force microscope is defined to be less than or equal to a predetermined value is useful for improving the reflectance of EUV light after the formation of a reflective film, and thus completed the present invention.

[0018] That is, the present invention is,

[0019] 1. A spatial frequency of 0.1 µm obtained by measuring the surface morphology of a 10 µm × 10 µm region using an atomic force microscope -1 20㎛ or more -1 The maximum value of the annular average power spectral density below is 1000 nm 4 Glass substrate for mask blanks characterized by the following:

[0020] 2. The above annular average power spectral density as af is a function of spatial frequency (f). -β In the case represented as such, the spatial frequency (f) is 1㎛ -1 10㎛ or more -1 A glass substrate for mask blanks of 1 having a coefficient β (fractal coefficient) of 0.7 or greater as described below,

[0021] 3. A glass substrate for mask blanks of 1 or 2, wherein the value of the surface roughness (RMS) obtained by measuring the surface shape of a 10㎛×10㎛ region with an atomic force microscope is 0.15 nm or less,

[0022] 4. A spatial frequency of 0.4 mm obtained by measuring the surface shape of a 6 mm × 6 mm area with a white interferometer -1 More than 100 mm -1 The maximum value of the below phantom average power spectral density is 10 12 nm 4 A glass substrate for mask blanks of any one of 1 to 3 below,

[0023] 5. A glass substrate for mask blanks of any one of 1 to 4, wherein the flatness (TIR) ​​of an area of ​​142 mm × 142 mm is 100 nm or less,

[0024] 6. A glass substrate for mask blanks of any one of 1 to 5, wherein the glass substrate for mask blanks is a titania-doped synthetic quartz glass substrate containing 5 to 10 mass% of titania.

[0025] It provides Effects of the invention

[0026] The glass substrate for mask blanks of the present invention has the characteristics of being highly flat, having a low defect, having a surface with low surface roughness, and having a sufficiently high reflectivity of EUV light after the formation of a reflective film, making it suitable as a substrate for EUVL mask blanks. Specific details for implementing the invention

[0027] The present invention will be described in more detail below.

[0028] The glass substrate for mask blanks of the present invention is used as a glass substrate for semiconductors when performing lithography technology that creates fine drawings using EUV light.

[0029] The size of the glass substrate for mask blanks is not particularly limited and can be of any size. For use in current EUVL exposure devices, a square substrate with sides of 6 inches, which is used for conventional photomask blank substrates, is preferred, and examples include a square 6025 substrate of 152 mm × 152 mm × 6.35 mm.

[0030] In addition, although the material is not particularly limited, in the photolithography process in EUVL, it is necessary to use a substrate with a small coefficient of thermal expansion, so a titania-doped synthetic quartz glass doped with titania at a concentration of 5 to 10 mass% is preferred.

[0031] The glass substrate for mask blanks of the present invention has a spatial frequency of 0.1 μm, obtained by measuring the surface shape of a 10 μm × 10 μm region on the substrate with an atomic force microscope in order to exhibit sufficient reflectance for EUV light after the formation of a multilayer reflective film. -1 20㎛ or more -1 The maximum value of the annular average power spectral density below is 1000 nm 4 It is characterized by the following.

[0032] In addition, the annular average power spectral density PSD(f), which is a function of the spatial frequency f, is calculated from the discrete Fourier transform F(u, v) of the surface shape Z(Px, Py) of the glass substrate. F(u, v) is calculated by the following equation (1).

[0033]

[0034] Here, Nx and Ny are measurement points in the x and y directions when measuring the surface shape of a glass substrate. Px and Py are integers representing the x and y positions of each measurement point, and take values ​​of Px = 0, 1, …, Nx - 1, and Py = 0, 1, …, Ny - 1. In contrast, u and v take values ​​of u = -1 / 2, -1 / 2 + 1 / Nx, …, 1 / 2, and v = -1 / 2, -1 / 2 + 1 / Ny, …, 1 / 2.

[0035] By normalizing F(u, v) with the measurement pitch Δx, Δy in the x and y directions and the area of ​​the measurement region A = (NxΔx) × (NyΔy) as in the following equation (2), the power spectral density P(u, v) can be obtained.

[0036]

[0037] Without this standardization, power spectral densities calculated from different measurement ranges or measurement pitch conditions cannot be simply compared.

[0038] Meanwhile, the spatial frequency f(u, v) is expressed by the following equation (3).

[0039]

[0040] The circular average power spectral density PSD(f) is the power spectral density P(u, v) averaged with respect to the spatial frequency f(u, v) as shown in the following equation (4).

[0041]

[0042] Here, Nf is the number of measurement points satisfying the following equation (5).

[0043]

[0044] Δf is defined by the following equation (6) if NxΔx=NyΔy.

[0045]

[0046] The maximum value of the above annular average power spectral density PSD(f) is at 1000 nm. 4 It can be less than or equal to this, but considering the need to further increase the reflectance to EUV light after the deposition of the multilayer reflective film, 750 nm 4 The following is desirable.

[0047] In addition, as an atomic force microscope, one can be appropriately selected from conventionally known ones, and specific examples thereof include the Oxford Instruments Cypher ES.

[0048] In addition, in the glass substrate for the mask blanks, the annular average power spectral density is af as a function of spatial frequency (f). -β In the case represented as such, the spatial frequency (f) is 1㎛ -1 10㎛ or more -1The coefficient β (fractal coefficient) below is preferably 0.7 or higher, and more preferably 0.8 or higher. The curve of the annular average power spectral density PSD decays rapidly from the low-frequency side toward the high-frequency side as the value of the fractal coefficient increases. Therefore, by controlling the fractal coefficient within this range as well as the maximum value of the PSD, it becomes possible to reduce surface roughness over a wide frequency range, resulting in improved reflectivity for EUV light after the formation of a multilayer reflective film, and also effectively suppressing high-frequency roughness components that can become noise during pattern formation.

[0049] In addition, the value of surface roughness (RMS) obtained by measuring the surface shape of a 10㎛×10㎛ area using an atomic force microscope is preferably 0.15 nm or less, and more preferably 0.10 nm or less. Since PSD is the average of roughness components in all angular directions within the measurement area as a function of spatial frequency, surface roughness originating from specific surface structures, such as structures occurring only in specific angular directions, may not be reflected. Therefore, by controlling the value of RMS, which is an indicator representing the surface roughness of the entire measurement field, within a range, the influence of surface structures that cause a decrease in reflectance to EUV light after the formation of a multilayer reflective film can be further eliminated.

[0050] As mentioned above, in titania-doped synthetic quartz glass substrates, the occurrence of a height-direction periodic structure due to the titanium concentration distribution can worsen the flatness of the substrate and reduce the precision of the exposure process; therefore, to achieve a higher precision exposure process, it is desirable to manage the PSD over a wider measurement field of view.

[0051] Accordingly, in the glass substrate for mask blanks of the present invention, the spatial frequency 0.4 mm obtained by measuring the surface shape of a 6 mm × 6 mm area by a white interferometer -1 More than 100 mm -1The maximum value of the below phantom average power spectral density is 10 12 nm 4 The following is preferable. Also, for the same reason, the flatness (TIR) ​​of the 142 mm × 142 mm area is preferably 100 nm or less, and more preferably 50 nm or less.

[0052] In addition, as a white interferometer, one can be appropriately selected from conventionally known ones, and specific examples thereof include the Nex View manufactured by Zygo.

[0053] The glass substrate for mask blanks of the present invention can be manufactured by a method comprising a rough polishing process for rough polishing a raw material substrate obtained by forming, annealing, slicing, chamfering, and lapping a glass ingot of the material, a flatness measuring process for measuring the flatness of the surface of the rough polished substrate, a partial polishing process for partial polishing, a finishing polishing process for finishing polishing, and a cleaning process for cleaning with a chemical solution.

[0054] The rough polishing process can be carried out by a double-sided polisher that performs planetary motion, using, for example, a cerium oxide-based abrasive as the abrasive.

[0055] Following the roughing process, the flatness of the substrate surface is measured in the flatness measurement process, but in order to reduce the processing time of the subsequent partial polishing process, it is preferable that the flatness (TIR) ​​of the 142 mm × 142 mm area be in the range of 100 to 1000 nm at the time after the roughing process is completed. The flatness measurement can be performed using a commercially available flatness tester for photomasks, such as the UltraFlat manufactured by Tropel.

[0056] In the partial polishing process, it is preferable to employ a partial polishing technique using a small rotary machining tool to create the shape of the glass substrate for the mask blanks of the present invention. In the flatness measurement process, the amount of polishing at each part of the substrate surface is determined based on the measurement data of the substrate surface measured in advance, and partial polishing is performed toward a preset target shape. The amount of polishing is controlled according to the speed at which the tool moves. That is, if a large amount of polishing is desired, the speed at which the tool passes through the substrate surface is slowed down; conversely, if the shape is already close to the target shape and a small amount of polishing is required, the amount of polishing is controlled by increasing the speed at which the tool passes through the substrate surface.

[0057] The machining part of the small rotary tool for partial grinding is not particularly limited, but it is preferable to use a rotary tool of the lute type.

[0058] Here, in order to reduce abrasive damage to the glass, the material of the rotary tool in contact with the glass can be selected from polyurethane, felt buff, rubber, cerium pad, etc., with a hardness of A50 to 75 (according to JISK 6253), but is not limited to these types as long as it can grind the glass surface.

[0059] In addition, the shape of the grinding part of the rotary machining tool is not particularly limited, and examples include circular, donut-shaped, cylindrical, bullet-shaped, disc-shaped, barrel-shaped, etc.

[0060] The flatness (TIR) ​​of the 142 mm × 142 mm area of ​​the substrate surface after the partial polishing process is preferably 100 nm or less, and more preferably 50 nm or less, and the shape can be arbitrarily selected according to specifications such as convex or concave depending on the conditions of the final finishing polishing.

[0061] In the finishing polishing process, batch polishing is performed on the substrate after the partial polishing process using conventional single-wafer polishing to improve defects or surface roughness that occurred up to the partial polishing process. At this time, a suede polishing cloth is suitably used. In addition, if the polishing rate is high, the shape created by partial polishing changes abruptly into the final target shape, and since the shape becomes difficult to control, it is preferable that the polishing rate is not very high.

[0062] In the finishing polishing process, finishing polishing is performed using abrasive particles. As these abrasive particles, a colloidal silica aqueous dispersion having an average primary particle size of 10 to 50 nm, preferably 10 to 20 nm, and a degree of aggregation of 1.0 to 1.8, preferably 1.0 to 1.3, is suitably used. If the average primary particle size is smaller than 10 nm, it becomes difficult to remove the abrasive particles from the surface of the glass substrate after polishing, which leads to an increase in microdefects; if the average primary particle size is larger than 50 nm, polishing marks caused by the abrasive particles are clearly left on the surface of the glass substrate, making it difficult to obtain the surface roughness required for the surface of the glass substrate for EUVL mask blanks. The degree of aggregation is defined as the ratio of the average secondary particle size to the average primary particle size, and since secondary particles refer to a collection (aggregation) of multiple primary particles, the degree of aggregation never becomes less than 1. When the degree of aggregation is greater than 1.8, the proportion of abrasive grains with anisotropy in shape increases, and non-uniformity occurs in the surface shape of the glass substrate after polishing, making it difficult to obtain the surface roughness required for the surface of glass substrates for EUVL mask blanks. Accordingly, by performing finish polishing with colloidal silica having controlled particle size and degree of aggregation, the spatial frequency in the 10㎛ × 10㎛ or smaller region required for the surface of glass substrates for EUVL mask blanks is 0.1㎛ -1 20㎛ or more -1For the following values ​​of the annular average power spectral density, it can be finalized within a suitable range.

[0063] In addition, the average primary particle size is calculated from the specific surface area of ​​the abrasive grains measured by the BET method, and the degree of association is calculated by dividing the average secondary particle size measured by the dynamic light scattering method by the average primary particle size.

[0064] In the cleaning process, the glass substrate for mask blanks after the finishing polishing process is cleaned in a single-wafer cleaner in a cleaning line containing a cleaning tank of an acidic or alkaline etching solution, and then dried. The etching solution is not particularly limited, and examples of acidic solutions include hydrofluoric acid, and alkaline solutions include aqueous solutions of KOH or NaOH.

[0065] In addition, if the amount of etching by the etching solution is large, the surface roughness of the substrate surface deteriorates, and the spatial frequency is 0.1㎛ -1 20㎛ or more -1 Since the value of the below average power spectrum density also increases, the etching amount is preferably 0.01 nm or less, and more preferably 0.005 nm or less.

[0066] For a glass substrate for mask blanks after cleaning, a Mo / Si periodic stacked film is formed by alternately stacking Mo films and Si films, for example. In this case, in order to form a multilayer reflective film suitable for EUV light of a wavelength of 13 to 14 nm, it is preferable to alternately stack Mo films and Si films with a film thickness of about a few nm for about 40 to 60 periods.

[0067] Next, a protective film containing, for example, Ru is formed on the multilayer reflective film.

[0068] Such multilayer reflective films or protective films can be formed, for example, by magnetron sputtering or ion beam sputtering.

[0069] Furthermore, in EUVL, since the entire optical system consists of reflective optics, slight differences in reflectivity at each reflective surface are accumulated over the number of reflections. Therefore, from the perspective of throughput and manufacturing costs, it is very important to maximize the reflectivity at each reflective surface.

[0070] [Example]

[0071] The present invention will be explained more specifically below with reference to examples and comparative examples, but the present invention is not limited to the following examples.

[0072] [Example 1]

[0073] After lapping and rough polishing a titania-doped synthetic quartz glass substrate (6 inches) doped with 7 mass% of titania as the raw material, local polishing was performed to remove high parts using a felt buffing tool based on height data of the substrate surface measured by a flatness tester (UltraFlat, manufactured by Tropel).

[0074] Afterwards, final finishing polishing was performed using a soft suede polishing cloth and a colloidal silica aqueous dispersion with a SiO2 concentration of 25 mass% as an abrasive (average primary particle size 14 nm, degree of association 1.3).

[0075] After polishing is finished, the substrate surface is cleaned in a cleaning line containing a cleaning tank that etches the substrate surface by about 0.008 nm using an etching solution adjusted to pH 10 in a single-wafer cleaning machine, and then dried.

[0076] The surface morphology of the obtained glass substrate was measured in a 10㎛×10㎛ region using an atomic force microscope (Oxford Instruments Cypher ES), and the RMS value was 0.04 nm and the spatial frequency was 0.1㎛ -1 20㎛ or more -1 The maximum value of the annular average power spectral density in the following region is 753 nm. 4 , 1㎛ -1 10㎛ or more -1The coefficient β (fractal coefficient) below was 0.7.

[0077] In addition, the surface shape of a 6 mm × 6 mm area was measured using a white interferometer (Zygo Nex View, hereinafter the same), and the spatial frequency was 0.4 mm -1 More than 100 mm -1 The maximum value of the annular average power spectral density below is 0.3×10 12 nm 4 It was. Also, the flatness (TIR) ​​of 142 mm × 142 mm was 28 nm.

[0078] Next, a substrate having a multilayer reflective film was fabricated by forming a multilayer reflective film (film thickness 300 nm) in which Mo films and Si films are alternately stacked 60 times on the main surface of the glass substrate for the mask blanks described above, and a protective film containing Ru (film thickness 2.5 nm) by ion beam sputtering. The reflectance of EUV light (wavelength 13.5 nm, hereinafter the same) on the main surface of the substrate having the multilayer reflective film was measured and found to be 65.8%.

[0079] [Example 2]

[0080] A glass substrate for mask blanks was fabricated under the same conditions as in Example 1, except that after the final finishing polishing was completed, the substrate surface was cleaned in a cleaning line containing a cleaning tank that etched the substrate surface by about 0.005 nm using an etching solution adjusted to pH 9 in a single-wafer cleaner, and then dried.

[0081] For the obtained glass substrate, the surface morphology of the substrate was measured using an atomic force microscope in a 10 µm × 10 µm region in the same manner as in Example 1, and the RMS value was 0.04 nm and the spatial frequency was 0.1 µm. -1 20㎛ or more -1 The maximum value of the annular average power spectral density in the following region is 721 nm. 4 , 1㎛ -1 10㎛ or more -1 The coefficient β (fractal coefficient) below was 0.8.

[0082] In addition, the surface shape of a 6 mm × 6 mm area was measured using a white interferometer, and the spatial frequency was 0.4 mm -1 More than 100 mm -1 The maximum value of the annular average power spectral density below is 0.1×10 12 nm 4 It was. Also, the flatness (TIR) ​​of 142 mm × 142 mm was 27 nm.

[0083] Next, a substrate having a multilayer reflective film was fabricated by forming a multilayer reflective film and a protective film on the main surface of the glass substrate for the mask blanks described above under the same conditions as in Example 1. The reflectance of EUV light on the main surface of the substrate having the multilayer reflective film was measured and found to be 66.0%.

[0084] [Comparative Example 1]

[0085] Glass substrates for mask blanks were fabricated under the same conditions as in Example 1, except that a colloidal silica aqueous dispersion with a SiO2 concentration of 25 mass% (average primary particle size 28 nm, degree of association 1.7) was used as the abrasive for the final finishing polishing, and a cleaning process using hydrofluoric acid to etch by 0.5 nm was added.

[0086] For the obtained glass substrate, the surface morphology of the substrate was measured using an atomic force microscope in a 10㎛ × 10㎛ region in the same manner as in Example 1, and the RMS value was 0.07 nm and the spatial frequency was 0.1㎛. -1 20㎛ or more -1 The maximum value of the annular average power spectral density in the following region is 1172 nm. 4 , 1㎛ -1 10㎛ or more -1 The coefficient β (fractal coefficient) below was 0.6.

[0087] In addition, the surface shape of a 6 mm × 6 mm area was measured using a white interferometer, and the spatial frequency was 0.4 mm -1 More than 100 mm -1The maximum value of the annular average power spectral density below is 1.2×10 12 nm 4 It was. Also, the flatness (TIR) ​​of 142 mm × 142 mm was 29 nm.

[0088] Next, a substrate having a multilayer reflective film was fabricated by forming a multilayer reflective film and a protective film on the main surface of the glass substrate for the mask blanks described above under the same conditions as in Example 1. The reflectance of EUV light on the main surface of the substrate having the multilayer reflective film was measured and found to be 64.2%.

[0089] In Comparative Example 1, final finishing polishing was performed using an abrasive with a relatively large particle size to increase the amount of etching in the cleaning process, and as a result, the surface morphology of the glass substrate for mask blanks obtained was 0.1 μm -1 20㎛ or more -1 The maximum value of the annular average power spectral density in the following region is 1172 nm 4 It grew larger.

[0090] As a result, the reflectance of the substrate having a multilayer reflective film obtained by depositing it on the glass substrate described above is 64.2%, which is a low value. When a pattern is exposed using mask blanks obtained from such a substrate having a multilayer reflective film, the low reflectance causes the exposure to take time, and there is concern that the throughput will decrease.

[0091] [Comparative Example 2]

[0092] A glass substrate for mask blanks was prepared under the same conditions as in Example 1, except that a colloidal silica aqueous dispersion with a SiO2 concentration of 25 mass% (average primary particle size 28 nm, degree of association 1.7) was used as the abrasive for the final finishing polishing.

[0093] For the obtained glass substrate, the surface morphology of the substrate was measured using an atomic force microscope in a 10 µm × 10 µm region in the same manner as in Example 1, and the RMS value was 0.06 nm and the spatial frequency was 0.1 µm. -1 20㎛ or more-1 The maximum value of the annular average power spectral density in the following region is 1035 nm. 4 , 1㎛ -1 10㎛ or more -1 The coefficient β (fractal coefficient) below was 0.6.

[0094] In addition, the surface shape of a 6 mm × 6 mm area was measured using a white interferometer, and the spatial frequency was 0.4 mm -1 More than 100 mm -1 The maximum value of the annular average power spectral density below is 1.1×10 12 nm 4 It was. Also, the flatness (TIR) ​​of 142 mm × 142 mm was 29 nm.

[0095] Next, a substrate having a multilayer reflective film was fabricated by forming a multilayer reflective film and a protective film on the main surface of the glass substrate for the mask blanks described above under the same conditions as in Example 1. The reflectance of EUV light on the main surface of the substrate having the multilayer reflective film was measured and found to be 64.4%.

[0096] In Comparative Example 2, as in Comparative Example 1, final finishing polishing was performed using an abrasive with a relatively large particle size; although the amount of etching during the cleaning process was suppressed, consequently, on the surface of the glass substrate for mask blanks obtained, 0.1 μm -1 20㎛ or more -1 The maximum value of the annular average power spectral density in the following region is 1035 nm. 4 It ended up becoming relatively large.

[0097] As a result, the reflectance of the substrate having a multilayer reflective film obtained by forming a film on the glass substrate described above is 64.4%, which is a relatively low value, and when mask blanks obtained from such a substrate are used, there is a concern that the throughput during exposure will decrease, similar to Comparative Example 1.

[0098] In addition, when comparing Comparative Example 2 and Example 1, the difference in reflectance is 1.4 percentage points, which may seem small; however, as mentioned above, in an EUVL optical system composed entirely of reflective optical systems, it is very important to maximize the reflectance at each reflective surface. In an exposure apparatus with 10 reflections within the system, when comparing the case where the reflectance of each reflective surface is 64.4% (Comparative Example 2) with the case where it is 65.8% (Example 1), the amount of EUV light on the exposure surface is 1.24 times higher in the latter case, and when comparing the case where the reflectance of each reflective surface is 64.2% (Comparative Example 1) with the case where it is 66.0% (Example 2), the amount of EUV light on the exposure surface is 1.32 times higher in the latter case, and these differences cannot be ignored in terms of throughput or manufacturing costs.

Claims

Claim 1 A spatial frequency of 0.1 µm obtained by measuring the surface morphology of a 10 µm × 10 µm region using an atomic force microscope -1 20㎛ or more -1 The maximum value of the annular average power spectral density below is 1000 nm 4 A glass substrate for mask blanks characterized by the following: Claim 2 A glass substrate for mask blanks according to claim 1, wherein the value of the surface roughness (RMS) obtained by measuring the surface shape of a 10㎛×10㎛ region with an atomic force microscope is 0.15 nm or less. Claim 3 In claim 1 or 2, a spatial frequency of 0.4 mm obtained by measuring the surface shape of a 6 mm × 6 mm area with a white interferometer. -1 More than 100 mm -1 The maximum value of the below phantom average power spectral density is 10 12 nm 4 Lee Ha-in, glass substrate for mask blanks. Claim 4 A glass substrate for mask blanks according to claim 1 or 2, wherein the flatness (TIR) ​​of a 142 mm × 142 mm area is 100 nm or less. Claim 5 A glass substrate for mask blanks according to claim 1 or 2, wherein the glass substrate for mask blanks is a titania-doped synthetic quartz glass substrate containing 5 to 10 mass% of titania. Claim 6 delete

Images