Method for manufacturing substrate for mask blank, method for manufacturing substrate with multilayer reflective film, method for manufacturing mask blank, method for manufacturing transfer mask, method for manufacturing semiconductor device, and substrate for mask blank

Abstract

Provided are a method for manufacturing a mask blank substrate, the surface of which is flat during actual use, and a mask blank substrate.　This method for manufacturing a mask blank substrate composed of a substrate having a first main surface and a second main surface opposite to the first main surface comprises a first measurement step (S10) for measuring the surface shape of the first main surface at least in an effective region, a second measurement step (S12) for measuring the surface shape of the second main surface at least in the effective region, an average virtual surface shape calculation step (S14) for calculating an average virtual surface shape having a secondary component in the substrate as a maximum order component at least in the effective region on the basis of the surface shapes obtained by measurement in the first measurement step and the second measurement step, and a local processing step (S16) for locally processing the first main surface and the second main surface on the basis of the average virtual surface shape.

Description

Method for manufacturing a mask blank substrate, method for manufacturing a substrate with a multilayer reflective film, method for manufacturing a mask blank, method for manufacturing a transfer mask, method for manufacturing a semiconductor device, and mask blank substrate

[0001] The present invention relates to a method for manufacturing a mask blank substrate comprising a substrate having a first main surface and a second main surface opposite to the first main surface, a method for manufacturing a substrate with a multilayer reflective film using a mask blank substrate, a method for manufacturing a mask blank, a method for manufacturing a transfer mask, a method for manufacturing a semiconductor device, and a mask blank substrate.

[0002] In recent years, EUV lithography, which uses EUV (Extreme Ultra Violet) light with a central wavelength of approximately 13.5 nm as a light source, has been investigated to further miniaturize semiconductor devices. Due to the characteristics of EUV light, EUV exposure requires a reflective optical system and a reflective mask. A reflective mask has a multilayer reflective film that reflects EUV light formed on a substrate, and an absorber film that absorbs EUV light is patterned on the multilayer reflective film. EUV light incident on the reflective mask from the illumination optical system of the exposure tool is reflected by areas without the absorber film (openings) and absorbed by areas with the absorber film (non-openings). The pattern on the absorber film becomes the mask pattern (transfer pattern). As a result, the mask pattern (transfer pattern) is transferred to a resist film on a semiconductor wafer as a resist pattern through the reduced projection optical system of the exposure tool, and subsequent processing is performed.

[0003] Reflective mask blanks used in the manufacture of reflective masks are required to have high flatness in order to improve the accuracy of the mask pattern (transfer pattern). Therefore, high flatness is also required for mask blank substrates. Therefore, for example, Patent Document 1 proposes a method for manufacturing a mask blank substrate consisting of a substrate having a main surface. The manufacturing method in Patent Document 1 includes a polishing step for polishing the main surface of the substrate, a first surface shape measurement step for measuring the surface shape of the main surface of the substrate at least in an effective area, a virtual surface shape calculation step for calculating, at least in the effective area, a virtual surface shape having, as a highest-order component, a second-order component corresponding to the surface unevenness shape measured in the first surface shape measurement step, and a local processing step for locally processing the main surface of the substrate based on the virtual surface shape.

[0004] Patent No. 6727879

[0005] When a reflective mask is used to expose a pattern onto a resist film or the like, the reflective mask is fixed to the flat surface of an electrostatic chuck for use. Therefore, the reflective mask needs to have a flat surface when fixed to the flat surface of the electrostatic chuck, which is the actual state of use. Patent Document 1 describes a method for manufacturing a mask blank substrate having a highly flat main surface, but does not take into account the actual state of use described above. Therefore, a reflective mask manufactured using a mask blank substrate manufactured by the manufacturing method described in Patent Document 1 does not guarantee surface flatness when fixed to the flat surface of the electrostatic chuck described above. The present invention aims to provide a method for manufacturing a mask blank substrate, a method for manufacturing a substrate with a multilayer reflective film, a method for manufacturing a mask blank, a method for manufacturing a transfer mask, a method for manufacturing a semiconductor device, and a mask blank substrate that have a flat surface in actual use.

[0006] As a result of extensive investigations, the present inventors have found that the above-mentioned problems can be solved by the following configuration: (1) A method for manufacturing a mask blank substrate comprising a substrate having a first main surface and a second main surface opposite to the first main surface, the method comprising: a first measurement step of measuring the surface shape of the first main surface at least in an effective area; a second measurement step of measuring the surface shape of the second main surface at least in the effective area; an average virtual surface shape calculation step of calculating an average virtual surface shape having a second-order component in the substrate as a maximum-order component at least in the effective area based on the surface shapes measured in the first measurement step and the second measurement step; and a local processing step of locally processing the first main surface and the second main surface based on the average virtual surface shape. (2) The method for manufacturing a mask blank substrate described in (1), wherein the average virtual surface shape calculation step is a step of obtaining first virtual surface shape data having a second-order component as the maximum order component from the surface shape of the first main surface obtained in the first measurement step, obtaining second virtual surface shape data having a second-order component as the maximum order component from the surface shape of the second main surface obtained in the second measurement step, and averaging the first virtual surface shape data and the second virtual surface shape data to obtain an average virtual surface shape.

[0007] (3) The method for manufacturing a mask blank substrate according to (1) or (2), wherein the average virtual surface shape calculation step is a step of obtaining first main surface shape data from the surface shape of the first main surface obtained in the first measurement step, obtaining second main surface shape data from the surface shape of the second main surface obtained in the second measurement step, averaging the first main surface shape data and the second main surface shape data to obtain average surface shape data, and obtaining an average virtual surface shape having a second-order component as the highest-order component from the average surface shape data. (4) The method for manufacturing a mask blank substrate according to any one of (1) to (3), wherein the method for calculating the average virtual surface shape in the virtual surface shape calculation step is the least squares method. (5) A method for manufacturing a mask blank substrate according to any one of (1) to (4), wherein, after finish polishing following the local processing step, the PV value in a BOW differential shape represented by the difference between a first main surface virtual surface shape having a second-order component of the first main surface of the substrate as its maximum order component and a second main surface virtual surface shape having a second-order component of the second main surface of the substrate as its maximum order component is 50 nm or less. (6) The method for manufacturing a mask blank substrate according to any one of (1) to (5), wherein the local processing step includes a step of translating an average virtual surface having an average virtual surface shape in the thickness direction of the substrate to set a first processing reference plane at a point in the unevenness of the first main surface of the substrate closest to the average virtual surface, and performing local processing to remove a region surrounded by the first main surface and the first processing reference plane, and a step of translating an average virtual surface having an average virtual surface shape in the thickness direction of the substrate to set a second processing reference plane at a point in the unevenness of the second main surface of the substrate closest to the average virtual surface, and performing local processing to remove a region surrounded by the second main surface and the second processing reference plane. (7) The method for manufacturing a mask blank substrate according to any one of (1) to (6), wherein the local processing uses ion beam etching. (8) The method for manufacturing a mask blank substrate according to any one of (1) to (7), wherein the measurement regions in the first measurement step and the second measurement step coincide when viewed from a direction perpendicular to the first main surface of the substrate.

[0008] (9) A method for manufacturing a multilayer reflective film-coated substrate, comprising the step of forming a multilayer reflective film on a first main surface of a mask blank substrate obtained by the method for manufacturing a mask blank substrate according to any one of (1) to (8). (10) A method for manufacturing a mask blank, comprising the step of forming an absorbing film on the first main surface of a mask blank substrate obtained by the method for manufacturing a mask blank substrate according to any one of (1) to (8), or on the multilayer reflective film of a multilayer reflective film-coated substrate obtained by the method for manufacturing a multilayer reflective film-coated substrate according to (9). (11) A method for manufacturing a transfer mask, comprising patterning the absorbing film of the mask blank obtained by the method for manufacturing a mask blank according to (10) to form a transfer pattern. (12) A method for manufacturing a semiconductor device, comprising the step of exposing and transferring a transfer pattern to a resist film on a semiconductor substrate using a transfer mask obtained by the method for manufacturing a transfer mask according to (11). (13) A mask blank substrate comprising a substrate having a first main surface and a second main surface opposite to the first main surface, wherein the PV value in a BOW differential shape expressed by the difference between a first main surface virtual surface shape having a second-order component of the first main surface of the substrate as its maximum order component and a second main surface virtual surface shape having a second-order component of the second main surface of the substrate as its maximum order component is 50 nm or less after finish polishing.

[0009] According to the present invention, it is possible to provide a method for manufacturing a mask blank substrate, a method for manufacturing a multilayer reflective film-coated substrate, a method for manufacturing a mask blank, a method for manufacturing a transfer mask, a method for manufacturing a semiconductor device, and a mask blank substrate, which have a flat surface in actual use.

[0010] FIG. 1 is a schematic cross-sectional view showing an example of a mask blank according to an embodiment of the present invention. FIG. 2 is a schematic perspective view showing an example of a substrate used in a method for manufacturing a mask blank substrate according to an embodiment of the present invention. FIG. 3 is a schematic view showing an example of a manufacturing system used in a method for manufacturing a mask blank substrate according to an embodiment of the present invention. FIG. 4 is a flowchart showing, in order of steps, an example of a method for manufacturing a mask blank substrate according to an embodiment of the present invention. FIG. 5 is a schematic view for explaining an example of an average virtual surface shape calculation step in a method for manufacturing a mask blank substrate according to an embodiment of the present invention. FIG. 6 is a schematic view for explaining another example of an average virtual surface shape calculation step in a method for manufacturing a mask blank substrate according to an embodiment of the present invention. FIG. 7 is a schematic view for explaining an example of a local processing step in a method for manufacturing a mask blank substrate according to an embodiment of the present invention. FIG. 8 is a schematic view for explaining an example of a local processing step in a method for manufacturing a mask blank substrate according to an embodiment of the present invention. FIG. 9 is a schematic view showing a state after a finish polishing step in a method for manufacturing a mask blank substrate according to an embodiment of the present invention.

[0011] Hereinafter, embodiments of the present invention will be described with reference to the drawings. However, the following embodiments are merely illustrative for explaining the present invention, and the present invention is not limited to the following embodiments. Various modifications and substitutions can be made to the following embodiments without departing from the scope of the present invention. Hereinafter, a numerical range expressed using "to" means a range including the numerical values ​​before and after "to" as the lower and upper limits. Unless otherwise specified, "parallel" and "perpendicular" include a generally accepted error range in the relevant technical field. Unless otherwise specified, "same" also includes a generally accepted error range in the relevant technical field.

[0012] A feature of the method for manufacturing a mask blank substrate of the present invention is a method for manufacturing a mask blank substrate comprising a substrate having a first main surface and a second main surface opposite to the first main surface, the method comprising: a first measurement step of measuring the surface shape of the first main surface at least in an effective area; a second measurement step of measuring the surface shape of the second main surface at least in the effective area; an average virtual surface shape calculation step of calculating an average virtual surface shape having a second-order component in the substrate as a maximum-order component at least in the effective area based on the surface shapes measured in the first and second measurement steps; and a local processing step of locally processing the first and second main surfaces based on the average virtual surface shape. In this way, the method for manufacturing a mask blank substrate calculates the average virtual surface shape having the second-order component in the substrate as a maximum-order component, and locally processes the first and second main surfaces based on the average virtual surface shape, thereby making the surfaces flat in an actual use state, i.e., when the substrate is fixed to the flat surface of an electrostatic chuck.

[0013] <Example of Mask Blank> Fig. 1 is a schematic cross-sectional view showing an example of a mask blank according to an embodiment of the present invention. The mask blank 10 shown in Fig. 1 is a reflective mask blank. The mask blank 10 is formed by laminating a conductive film 19, a substrate 12, a multilayer reflective film 14, a protective film 16, and an absorber film 18 in this order. For example, the main component of the substrate 12 is SiO ２ 1 may have no protective film 16. The mask blank 10 may also have a protective film (not shown), which will be described later, on the side of the conductive film 19 opposite to the substrate 12 side. The mask blank 10 is not limited to a reflective mask blank, and may also be a transmissive mask blank.

[0014] (Mask Blank Substrate) FIG. 2 is a schematic perspective view showing an example of a substrate used in a mask blank substrate manufacturing method according to an embodiment of the present invention. The mask blank substrate 13 shown in FIG. 2 will eventually become the substrate 12 of the mask blank 10. The mask blank substrate 13 is composed of a substrate 15 having a first main surface 15a and a second main surface 15b opposite the first main surface 15a. The mask blank substrate 13 and the substrate 15 are, for example, plate-like members having a rectangular shape in a planar view. In the mask blank substrate 13 and the substrate 15, the first main surface 15a and the second main surface 15b have the same shape and size and have side surfaces 15c surrounding the first main surface 15a and the second main surface 15b. Each of the first main surface 15a and the second main surface 15b has an effective area. The effective area is an area that includes at least an area defined as a quality area. The quality area is determined by a standard. When the mask blank substrate 13 (substrate 15) has a substrate size of 152 mm × 152 mm, the quality area is 132 mm × 104 mm for the first main surface 15 a (front surface) and 138 mm × 104 mm for the second main surface 15 b (rear surface). When the substrate is a glass substrate, taking into consideration the management of the front and back surfaces or orientation of the glass substrate, the effective area of ​​the first main surface 15 a and the second main surface 15 b, i.e., both the front and rear surfaces, is preferably 138 mm × 138 mm or larger, and may be 142 mm × 142 mm or larger. Alignment marks (not shown) may be provided on the first main surface 15 a and the second main surface 15 b of the substrate 15 to align the measurement positions of the respective surfaces.

[0015] <Example of Manufacturing System> FIG. 3 is a schematic diagram showing an example of a manufacturing system used in the mask blank substrate manufacturing method according to an embodiment of the present invention. The manufacturing system 20 includes a measurement unit 22, a calculation unit 24, a processing unit 26, a finish polishing unit 28, and a display unit 29. The measurement unit 22 measures the surface shapes of the mask blank substrate 13 and the substrate 15. The measurement unit 22 also measures the surface shape of the substrate 15 after finish polishing. The measurement unit 22 measures the surface shape of the first main surface 15a at least in the effective area, and measures the surface shape of the second main surface 15b at least in the effective area. This results in first main surface shape data representing the surface shape of the first main surface 15a, and second main surface shape data representing the surface shape of the second main surface 15b. The measurement unit 22 includes, for example, an optical interferometer. The surface shape of the substrate 15 is measured by, for example, the optical interferometer. Examples of the optical interferometer include a Fizeau interferometer or a grazing incidence interferometer. The measured first principal surface shape data (not shown) representing the surface shape of the first principal surface 15a and the second principal surface shape data (not shown) representing the surface shape of the second principal surface 15b are output to the calculation unit 24 and stored in a memory (not shown) of the calculation unit 24. The first principal surface shape data (not shown) representing the surface shape of the first principal surface 15a and the second principal surface shape data (not shown) representing the surface shape of the second principal surface 15b are each acquired for each measurement point in the form of height information (convexoconcave shape information) relative to a virtual absolute plane, which will be described later. Therefore, various data, which will be described later, obtained using the first principal surface shape data representing the surface shape of the first principal surface 15a and the second principal surface shape data representing the surface shape of the second principal surface 15b are also expressed in the form of height information relative to a virtual absolute plane, which will be described later.

[0016] Surface shapes such as the first surface shape and the second surface shape are generally expressed by the following formula (1) using three orthogonal coordinates, x, y, and z (see FIGS. 5 and 6). Surface shape data is expressed by the numerical values ​​of the three orthogonal coordinates, x, y, and z. For the three orthogonal coordinates, x, y, and z, for example, x is the abscissa, y is the ordinate, and z is the height coordinate. A shape having a second-order component as the highest-order component is a shape expressed by formula (2), and a simplified formula is formula (3) for a quadratic curved surface shape (paraboloid). Here, aｉ，ｊ is a coefficient, and i and j are integers.

[0017]

[0018] z = a ０，０ +a １，０ ×x+a ０，１ ×y+a １，１ ×x×y + a ２，０ ×x ２ +a ０，２ xy ２ (2) z = a ２，０ ×x ２ +a ０，２ xy ２ (3) a, which is the zero-order component (height reference component) ０，０ , and the first-order component (linear gradient component) a １，０ ×x and a ０，１ When a transfer mask manufactured using a mask blank substrate is placed on an exposure tool and used, leveling adjustment (height adjustment) or tilt correction (inclination correction) is performed on the exposure tool side, so there is no problem if xy is included. In the following explanation, for simplicity, the zeroth-order and first-order components will be omitted. Note that, if the surface shape of the first principal surface acquired in the first measurement step (step S10 described below) for measuring the surface shape of the first principal surface is expressed as a concave-convex shape relative to a virtual absolute plane, the zeroth-order and first-order components are incorporated into this virtual absolute plane, which is preferable because it makes handling easier. Similarly, if the surface shape of the second principal surface acquired in the second measurement step (step S12 described below) for measuring the surface shape of the second principal surface is expressed as a concave-convex shape relative to a virtual absolute plane, the zeroth-order and first-order components are incorporated into this virtual absolute plane, which is preferable because it makes handling easier.

[0019] Here, the virtual absolute plane is a plane that has the smallest value when the root mean square of the distance from the reference plane to the main surface of the substrate is calculated for the entire flatness measurement region. The surface shape after local processing tends to be close to centrosymmetric about the center of the main surface of the substrate. Taking this into consideration, the virtual surface shapes such as the first virtual surface shape and the second virtual surface shape are expressed by Equation (3), which is centrosymmetric about the center of the main surface of the substrate. Note that if the central axis of the surface shape after polishing is eccentric from the center of the main surface of the substrate, the eccentricity can be taken into account when applying the formula. In other words, if the eccentric point from the center of the main surface of the substrate is (x ０ , y ０ ), then it corresponds to equation (4). ２，０ ×(x-x ０ ) ２ +a ０，２ ×(y-y ０ ) ２ (4) From the above, a virtual surface shape having a second-order component as the maximum order component is treated as a virtual quadratic curved surface shape (virtual paraboloid shape) expressed by equation (2).

[0020] The virtual surface shape is preferably a shape that minimizes the amount of processing when performing local processing on the main surface of the substrate, as described below. A method for calculating a virtual surface shape that reduces the amount of processing during local processing is the least squares method. Therefore, a curved surface (paraboloid) consisting of quadratic components expressed by equation (2) is fitted to the main surface of the substrate using the least squares method to obtain the coefficient a ２，０ and a ０，２ It is desirable to calculate the virtual surface shape using a curved surface (paraboloid) consisting of the calculated quadratic components. If the amount of processing is small, the time for local processing is short, manufacturing throughput is improved, processing errors are reduced, and the frequency of scratches during local processing is also reduced. Therefore, the amount of finish polishing can be reduced, and deformation from this virtual surface shape due to finish polishing is also reduced, making it easier to obtain the desired flatness.

[0021] The calculation unit 24 calculates an average virtual surface shape (average virtual surface shape data) at least in the effective area. The average virtual surface shape has a second-order component of the substrate 15 as the highest-order component based on the surface shapes of the first main surface 15a and the second main surface 15b. The method for calculating the average virtual surface shape is, for example, the least squares method. The average virtual surface shape will be described in detail later. The calculation unit 24 calculates the average virtual surface shape (average virtual surface shape data) using first main surface shape data representing the surface shape of the first main surface 15a and second main surface shape data representing the surface shape of the second main surface 15b. The calculation unit 24 may be configured, for example, by a computer equipped with an arithmetic processing unit such as a CPU (Central Processing Unit), or may be a dedicated device configured with a dedicated circuit, or may be configured as a server to be executed on the cloud. The calculation unit 24 calculates the average virtual surface shape by performing calculation processing using a program (computer software) stored in, for example, a ROM (Read Only Memory) or the like.

[0022] The processing unit 26 locally processes the first main surface 15a and the second main surface 15b of the substrate 15 based on the average virtual surface shape. The local processing is performed based on the processing amount of the first main surface and the processing amount of the second main surface, which will be described later. For example, ion beam etching is used for the local processing, but gas cluster ion beam etching is preferably used. The configuration of the processing unit 26 is not particularly limited as long as it can perform the above-mentioned ion beam etching. For example, in addition to ion beam etching, magnetorheological fluid polishing (MRF), which is a magnetic abrasive slurry processing method, and local chemical mechanical polishing (LCMP) can also be used for the local processing. The processing amount of the first main surface and the processing amount of the second main surface, which will be described later, are calculated, for example, by the calculation unit 24.

[0023] The finish polishing unit 28 performs finish polishing on the first main surface 15a and the second main surface 15b of the substrate 15 that have been locally processed by the processing unit 26. The finish polishing unit 28 uses, for example, a double-sided polishing machine or a single-sided polishing machine, and polishes using a polishing pad and a polishing slurry. The polishing slurry may be, for example, colloidal silica slurry. The finish polishing may be performed using a double-sided polishing machine to simultaneously polish the first main surface 15a and the second main surface 15b of the substrate 15. Alternatively, the first main surface and the second main surface may be polished sequentially using a single-sided polishing machine.

[0024] The display unit 29 is connected to the measurement unit 22, the calculation unit 24, the processing unit 26, and the finish polishing unit 28. The display unit 29 displays the surface shape of the first main surface, the surface shape of the second main surface, and the average virtual surface shape. The display unit 29 can also display the amount of local processing on the first main surface and the amount of local processing on the second main surface, as well as the area on the first main surface side and the area on the second main surface side of the local processing. The display unit 29 is not particularly limited, and various known displays such as a liquid crystal display can be used. Furthermore, the display screen of a mobile communication terminal such as a smartphone or tablet terminal can be used as the display unit 29.

[0025] <Example of Mask Blank Substrate Manufacturing Method> FIG. 4 is a flowchart showing an example of a mask blank substrate manufacturing method according to an embodiment of the present invention, in order of steps. The mask blank substrate manufacturing method will be described using, for example, the manufacturing system 20 shown in FIG. 2 . However, the mask blank substrate manufacturing method is not particularly limited to using the manufacturing system 20 shown in FIG. 2 . The mask blank substrate manufacturing method first performs a first measurement step (step S10) in which the surface shape of the first main surface 15 a of the substrate 15 that will become the mask blank substrate 13 shown in FIG. 2 is measured at least in the effective area, as shown in FIG. 4 . In step S10, the surface shape of the first main surface 15 a of the substrate 15 is measured using the measurement unit 22 of the manufacturing system 20, for example, an optical interferometer, to obtain the surface shape of the first main surface 15 a of the substrate 15. In step S10, shape data (first main surface shape data) of the surface shape of the first main surface 15 a of the substrate 15 is obtained and output to the calculation unit 24.

[0026] Next, a second measurement step is performed to measure the surface shape of the second main surface 15b of the substrate 15, which will become the mask blank substrate 13 shown in Fig. 2, at least in the effective area, as shown in Fig. 4 (step S12). In step S12, the measurement unit 22 of the manufacturing system 20, for example, an optical interferometer, is used to measure the surface shape of the second main surface 15b of the substrate 15, thereby obtaining the surface shape of the second main surface 15b of the substrate 15. In step S12, shape data of the surface shape of the second main surface 15b of the substrate 15 (second main surface shape data) is obtained and output to the calculation unit 24. Next, an average virtual surface shape calculation step is performed to calculate an average virtual surface shape of the substrate 15, which has a second-order component as the highest-order component, at least in the effective area, based on the surface shapes measured in the first measurement step (step S10) and the second measurement step (step S12) (step S14).

[0027] The average virtual surface shape calculation step (step S14) is performed, for example, by the calculation unit 24. In the average virtual surface shape calculation step (step S14), first virtual surface shape data having a second-order component as the highest-order component is obtained from the surface shape of the first principal surface 15a (first principal surface shape data of the surface shape) obtained in the first measurement step (step S10) using the least squares method. Also, second virtual surface shape data having a second-order component as the highest-order component is obtained from the surface shape of the second principal surface 15b (second principal surface shape data of the surface shape) obtained in the second measurement step (step S12) using, for example, the least squares method. Then, the first virtual surface shape data and the second virtual surface shape data are averaged to obtain average virtual surface shape data, thereby obtaining the average virtual surface shape. The average virtual surface shape is represented by the average virtual surface shape data. The average of the first virtual surface shape data and the second virtual surface shape data is not particularly limited, but may be, for example, a simple average (arithmetic mean). Here, the second-order component means that it is expressed by a quadratic function. That is, the average virtual surface shape, the first virtual surface shape data, and the second virtual surface shape data can all be expressed by quadratic functions. The first virtual surface shape data having a second-order component as its highest-order component and the second virtual surface shape data having a second-order component as its highest-order component are not particularly limited to being calculated using the least squares method.

[0028] 5 is a schematic diagram illustrating an example of an average virtual surface shape calculation step in the mask blank substrate manufacturing method according to an embodiment of the present invention. More specifically, as shown in FIG. 5 , in a first measurement step (step S10) of a substrate 15, a surface shape 30 of the first main surface 15a is obtained. The surface shape 30 is represented by first main surface shape data of the surface shape 30. From the surface shape 30, a first virtual surface shape 33 having a second-order component as its highest-order component is obtained, for example, using the least squares method. The first virtual surface shape 33 is represented by the first virtual surface shape data. In a second measurement step (step S12) of the substrate 15, a surface shape 32 of the second main surface 15b is obtained. The surface shape 32 is represented by second main surface shape data of the surface shape 32. From the surface shape 32, a second virtual surface shape 35 having a second-order component as its highest-order component is obtained, for example, using the least squares method. The second virtual surface shape 35 is represented by the second virtual surface shape data. Next, the first virtual surface shape data (first virtual surface shape 33) and the second virtual surface shape data (second virtual surface shape 35) are averaged to obtain average virtual surface shape data, and an average virtual surface shape 36 represented by the average virtual surface shape data is obtained.

[0029] The average virtual surface shape calculation step (step S14) is not particularly limited to the above-described steps. For example, first principal surface shape data is obtained from the surface shape of the first principal surface 15a obtained in the first measurement step (step S10), and second principal surface shape data is obtained from the surface shape of the second principal surface 15b obtained in the second measurement step (step S12). The first principal surface shape data and the second principal surface shape data are averaged to obtain average surface shape data. For example, the least squares method is used to obtain average virtual surface shape data having a second-order component as the highest-order component from the average surface shape data, thereby obtaining the average virtual surface shape. The average of the first principal surface shape data and the second principal surface shape data is not particularly limited, and may be, for example, a simple average (arithmetic mean).

[0030] 6 is a schematic diagram illustrating another example of an average virtual surface shape calculation step in the mask blank substrate manufacturing method according to an embodiment of the present invention. More specifically, as shown in FIG. 6 , in a first measurement step (step S10), a surface shape 30 of the first main surface 15 a of the substrate 15 is obtained. The surface shape 30 is represented by first main surface shape data of the surface shape 30. In a second measurement step (step S12), a surface shape 32 of the second main surface 15 b of the substrate 15 is obtained. The surface shape 32 is represented by second main surface shape data of the surface shape 32. Next, the first main surface shape data of the surface shape 30 of the first main surface 15 a and the first main surface shape data of the surface shape 32 of the second main surface 15 b are averaged to obtain average surface shape data. An average virtual surface shape 34 represented by the average surface shape data is obtained. For example, using the least squares method, average virtual surface shape data having a second-order component as the highest-order component is obtained from the average surface shape data representing the average virtual surface shape 34. An average virtual surface shape 36 represented by the average virtual surface shape data is obtained.

[0031] As shown in Fig. 4 , after the average virtual surface shape calculation step (step S14), a local machining step (step S16) is performed in which the first main surface 15a and the second main surface 15b of the substrate 15 shown in Fig. 2 are locally machined based on the average virtual surface shape. The local machining step (step S16) is performed by the machining unit 26. The local machining step (step S16) includes a step of setting the average virtual surface shape as a first machining reference plane for the first main surface 15a of the substrate 15 and performing local machining to remove regions of the first main surface 15a that are higher than the first machining reference plane, and a step of setting the average virtual surface shape as a second machining reference plane for the second main surface 15b of the substrate 15 and performing local machining to remove regions of the second main surface 15b that are higher than the second machining reference plane.

[0032] 7 and 8 are schematic diagrams illustrating an example of a local processing step in a mask blank substrate manufacturing method according to an embodiment of the present invention. More specifically, in the local processing step (step S16), an average virtual surface 37 having the average virtual surface shape 36 shown in FIGS. 5 and 6 is translated in the thickness direction of the substrate 15, and a first processing reference plane 36a is set at the location closest to the average virtual surface 37 in the unevenness of the first main surface 15a of the substrate 15, as shown in FIG. 7. That is, the average virtual surface 37 is fitted to the bottom of the recess located toward the center of the thickness direction of the substrate 15 among the unevenness of the first main surface 15a of the substrate 15, thereby setting the first processing reference plane 36a shown in FIG. 7. A region 38 surrounded by the first main surface 15a and the first processing reference plane 36a is the processing area to be removed. The calculation unit 24 calculates the amount of processing of the first main surface 15a based on the first processing reference plane 36a. Furthermore, an average virtual surface 37 having the average virtual surface shape 36 is translated in the thickness direction of the substrate 15, and a second machining reference plane 36b is set at, for example, the point of the unevenness of the second main surface 15b of the substrate 15 closest to the average virtual surface 37, as shown in FIG. 7. That is, the average virtual surface 37 is fitted to the bottom of the recess that is located toward the center of the thickness direction of the substrate 15 among the unevenness of the second main surface 15b of the substrate 15, and the second machining reference plane 36b shown in FIG. 7 is set. A region 39 surrounded by the second main surface 15b and the second machining reference plane 36b is the machining region to be removed. The calculation unit 24 calculates the amount of machining of the second main surface 15b based on the second machining reference plane 36b.

[0033] The processing unit 26 performs local processing using, for example, ion beam etching on the first main surface 15a and the second main surface 15b of the substrate 15, to remove a region 38 on the first main surface 15a side and a region 39 on the second main surface 15b side, based on the above-mentioned processing amount of the first main surface 15a and the processing amount of the second main surface 15b. As a result, for example, the mask blank substrate 13 shown in FIG. 8 is obtained.

[0034] After the local processing step (step S16), the substrate is subjected to a finish polishing step (step S18) using, for example, the processing unit 26. In the finish polishing step, for example, the first and second main surfaces are simultaneously polished using a double-sided polisher (not shown). However, the first and second main surfaces may be polished sequentially using a single-sided polisher. Furthermore, in the finish polishing step, the substrate is polished while a polishing slurry is supplied between the polishing pad and the substrate. The polishing slurry contains an abrasive such as colloidal silica particles. When the substrate 15 (mask blank substrate 13) after the finish polishing step (step S18) is brought into close contact with a horizontal plane Hp, as shown in FIG. 9 , assuming exposure, which is an actual use state, the first main surface 15a of the substrate 15 becomes flat. In this way, a mask blank substrate 13 with a flat surface is obtained in actual use. It is known that if the curvature of the substrate includes a second-order component (parabolic), it can be easily corrected by the exposure apparatus and can be treated as an optically flat surface during exposure. For this reason, the second-order component is set as the highest-order component, as described above. After finish polishing following the local processing step, the PV value of the BOW differential shape expressed by adding the first main surface virtual surface shape having the second-order component of the first main surface of the substrate as its highest-order component and the second main surface virtual surface shape having the second-order component of the second main surface of the substrate as its highest-order component is preferably 50 nm or less. The mask blank substrate 13 preferably has a PV value of 50 nm or less in the BOW differential shape after finish polishing. The PV value is a value indicating the difference between the highest point (peak) and the lowest point (valley) in the BOW differential shape.

[0035] The BOW differential shape is obtained as follows. The first main surface 15a of the substrate 15 after the local processing step shown in FIG. 8 is measured using, for example, an optical interferometer to obtain first main surface shape data of the surface shape of the first main surface 15a. Next, first virtual surface shape data having a second-order component as the highest-order component is obtained from the first main surface shape data of the surface shape using, for example, the least squares method. The second main surface 15b of the substrate 15 after the local processing step is measured using, for example, an optical interferometer to obtain second main surface shape data of the surface shape of the second main surface 15b. Next, second virtual surface shape data having a second-order component as the highest-order component is obtained from the second main surface shape data of the surface shape using, for example, the least squares method. Next, in order to align the orientation of the surface shape represented by the first virtual surface shape data with the surface shape represented by the second virtual surface shape data, for example, the second virtual surface shape data is flipped left and right to create inverted second virtual surface shape data. Next, the first virtual surface shape data and the inverted second virtual surface shape data are added together to obtain BOW differential shape data. This results in a BOW differential shape. In the BOW differential shape, the highest point (peak) and the lowest point (valley) correspond to the maximum and minimum values ​​of the BOW differential shape data, and the difference between the maximum and minimum values ​​of the BOW differential shape data is the above-mentioned PV value. Note that, since it is only necessary to align the orientations of the surface shape represented by the first virtual surface shape data and the surface shape represented by the second virtual surface shape data, it is also possible to create and use inverted first virtual surface shape data by inverting the first virtual surface shape data left and right without inverting the second virtual surface shape data left and right. If alignment marks are provided on the first main surface 15a and the second main surface 15b of the substrate 15, the above-mentioned first virtual surface shape data or second virtual surface shape data can be inverted left and right based on the alignment marks.

[0036] As shown in FIG. 9 , the substrate 15 after the finish polishing process following the local processing step, i.e., the mask blank substrate 13 after the finish polishing, has a flat first main surface 15a when the second main surface 15b of the substrate 15 is brought into close contact with a horizontal plane Hp, assuming exposure, which is an actual use state. There is an index value that indicates this. The index value is obtained as follows. First, the second main surface 15b of the substrate 15 is measured, for example, using an optical interferometer to obtain second main surface shape data of the surface shape of the second main surface 15b. Next, second virtual surface shape data having a second-order component as the highest-order component is obtained from the second main surface shape data of the surface shape, for example, using the least squares method. Next, to align the orientation of the surface shape represented by the first main surface shape data with the surface shape represented by the second virtual surface shape data, for example, the second virtual surface shape data is flipped left and right to create inverted second virtual surface shape data. Next, the first main surface shape data and the inverted second virtual surface shape data are added together to obtain index bow shape data. This results in an index bow shape. The highest and lowest points in the index bow shape correspond to the maximum and minimum values ​​of the index bow shape data, and the difference between the maximum and minimum values ​​of the index bow shape data is the above-mentioned index value. Note that, since it is only necessary to align the orientations of the surface shape represented by the first principal surface shape data and the surface shape represented by the second virtual surface shape data, it is also possible to create inverted first principal surface shape data by left-right inverting the first principal surface shape data without left-right inverting the second virtual surface shape data, and use this inverted first principal surface shape data.

[0037] It is preferable that the measurement areas in the first measurement step (step S10) and the second measurement step (step S12) coincide when viewed from a direction perpendicular to the first main surface 15a of the substrate 15 shown in FIG. 2. This improves the measurement accuracy of the BOW differential shape and the like. Note that "the measurement areas coincide" means that the measurement areas are not only coincident in size but also coincident in position. Furthermore, "coincidence" includes the error range acceptable in the technical field.

[0038] The method for manufacturing a mask blank substrate is not limited to manufacturing a mask blank substrate by performing the above-mentioned first measurement step (step S10), second measurement step (step S12), average virtual surface shape calculation step (step S14), and local processing step (step S16) once. For example, in the method for manufacturing a mask blank substrate, after the local processing step (step S16), a PV value in the BOW differential shape may be calculated as described above, and if the PV value is greater than a set value, the above-mentioned first measurement step (step S10), second measurement step (step S12), average virtual surface shape calculation step (step S14), and local processing step (step S16) may be repeatedly performed until the PV value becomes smaller than the set value.

[0039] <Method for manufacturing a substrate with a multilayer reflective film> A substrate with a multilayer reflective film can be manufactured using a mask blank substrate obtained by the method for manufacturing a mask blank substrate. The method for manufacturing a substrate with a multilayer reflective film includes a step of forming a multilayer reflective film on the first main surface of the mask blank substrate obtained by the method for manufacturing a mask blank substrate. The multilayer reflective film and the method for forming the multilayer reflective film will be described in detail later. Below, a reflective mask blank will be described in more detail.

[0040] (Reflective Mask Blank) [Mask Blank Substrate] The mask blank substrate preferably has a small thermal expansion coefficient. The mask blank substrate preferably has a thermal expansion coefficient of 0±1.0×10 at 20° C. －７ / °C, and 0±0.3×10 －７ / °C is more preferable. Examples of materials with a small thermal expansion coefficient include SiO ２ -TiO ２ However, the present invention is not limited to this, and substrates made of crystallized glass in which β-quartz solid solution is precipitated, quartz glass, metallic silicon, metal, and the like can also be used. ２ -TiO ２ The SiO-based glass ２ 90 to 95 mass% of TiO ２ It is preferable to use quartz glass containing 5 to 10 mass % of the above.

[0041] The surface of the substrate on which the multilayer reflective film is laminated (hereinafter also referred to as the "first principal surface") preferably has high surface smoothness. The surface smoothness of the first principal surface can be evaluated by surface roughness. The surface roughness of the first principal surface is preferably 0.15 nm or less in terms of root-mean-square roughness Rq. The surface roughness can be measured using an atomic force microscope and is described as the root-mean-square roughness Rq based on JIS (Japanese Industrial Standards)-B0601. The first principal surface is preferably surface-processed to achieve a predetermined flatness, in order to improve the pattern transfer accuracy and positional accuracy of a reflective mask obtained using the reflective mask blank. In a predetermined region of the first principal surface (e.g., a 132 mm × 132 mm region), the substrate preferably has a flatness of 100 nm or less, more preferably 50 nm or less, and even more preferably 30 nm or less. The flatness can be measured using a flatness measuring instrument manufactured by Fujifilm Corporation. The size and thickness of the substrate are determined appropriately depending on the design values ​​of the mask, etc. For example, the outer shape may be 6 inches (152 mm) square and the thickness may be 0.25 inches (6.3 mm). Furthermore, the substrate preferably has high rigidity in order to prevent deformation due to film stress of films (multilayer reflective film, absorber film, etc.) formed on the substrate. For example, the Young's modulus of the substrate is preferably 65 GPa or more.

[0042] [Multilayer reflective film] The multilayer reflective film provided on one surface of the reflective mask blank is not particularly limited as long as it has the desired properties as a reflective film for an EUV mask blank. The multilayer reflective film preferably has a high reflectivity for EUV light. Specifically, when EUV light is incident on the surface of the multilayer reflective film at an incident angle of 6°, the maximum reflectivity for EUV light with a wavelength of around 13.5 nm is preferably 60% or more, more preferably 65% ​​or more. Similarly, even when a protective film is laminated on the multilayer reflective film, the maximum reflectivity for EUV light with a wavelength of around 13.5 nm is preferably 60% or more, more preferably 65% ​​or more.

[0043] Since multilayer reflective films can achieve high reflectivity for EUV light, they are typically formed by alternately stacking multiple high-refractive index layers that exhibit a high refractive index for EUV light and multiple low-refractive index layers that exhibit a low refractive index for EUV light. The multilayer reflective film may be formed by stacking multiple cycles of a stack structure in which high-refractive index layers and low-refractive index layers are stacked in this order from the substrate side, or multiple cycles of a stack structure in which low-refractive index layers and high-refractive index layers are stacked in this order. The high-refractive index layer may be a layer containing Si. Examples of materials containing Si include elemental Si and Si compounds containing one or more elements selected from the group consisting of B, C, N, and O. The use of a high-refractive index layer containing Si results in a reflective mask with excellent reflectivity for EUV light. The low-refractive index layer may be a layer containing a metal selected from the group consisting of Mo, Ru, Rh, and Pt, or an alloy thereof. Si is commonly used for the high-refractive index layer, and Mo is commonly used for the low-refractive index layer. That is, Mo / Si multilayer reflective films are most common, but the multilayer reflective films are not limited to this, and Ru / Si multilayer reflective films, Mo / Be multilayer reflective films, Mo compound / Si compound multilayer reflective films, Si / Mo / Ru multilayer reflective films, Si / Mo / Ru / Mo multilayer reflective films, and Si / Ru / Mo / Ru multilayer reflective films can also be used.

[0044] The thickness of each layer constituting the multilayer reflective film and the number of repeating units of the layer can be appropriately selected depending on the film material used and the reflectivity of the reflective layer required for EUV light. Taking a Mo / Si multilayer reflective film as an example, to obtain a multilayer reflective film with a maximum reflectivity of 60% or more for EUV light, it is sufficient to laminate a Mo film with a thickness of 2.3±0.1 nm and a Si film with a thickness of 4.5±0.1 nm so that the number of repeating units is 30 to 60.

[0045] Each layer constituting the multilayer reflective film can be deposited to a desired thickness using a known deposition method, such as magnetron sputtering or ion beam sputtering. For example, when a multilayer reflective film is fabricated using ion beam sputtering, ion particles are supplied from an ion source to a target of a high refractive index material and a target of a low refractive index material. When the multilayer reflective film is a Mo / Si multilayer reflective film, for example, a Si layer having a predetermined thickness is first deposited on a substrate using an ion beam sputtering method. Then, a Mo layer having a predetermined thickness is deposited using an Mo target. This Si layer and Mo layer constitute one cycle, and 30 to 60 cycles are stacked to form a Mo / Si multilayer reflective film.

[0046] [Protective Film] The reflective mask blank may have a protective film on the side of the multilayer reflective film opposite the substrate side. The protective film is provided for the purpose of protecting the multilayer reflective film from damage during etching (usually a dry etching process) when a pattern is formed on the absorber film by the etching process. Examples of materials that can achieve the above purpose include materials containing at least one element selected from the group consisting of Ru and Rh. That is, it is preferable that the protective film contains at least one element selected from the group consisting of Ru and Rh. More specifically, the above-mentioned materials include Ru metal alone, Ru alloys containing Ru and one or more metals selected from the group consisting of Si, Ti, Nb, Mo, Rh, and Zr, Rh metal alone, Rh alloys containing Rh and one or more metals selected from the group consisting of Si, Ti, Nb, Ru, Ta, Mo, and Zr, Rh-containing nitrides containing the above-mentioned Rh alloys and nitrogen, and Rh-containing oxynitrides containing the above-mentioned Rh alloys, nitrogen, and oxygen. Also, materials that can achieve the above-mentioned object include Al and nitrides containing these metals and nitrogen, and Al. ２ O ３Among these, materials that can achieve the above-mentioned object are preferably Ru metal alone, Ru alloys, Rh metal alone, or Rh alloys. As the Ru alloy, a Ru—Si alloy or a Ru—Rh alloy is preferred, and as the Rh alloy, a Rh—Si alloy or a Rh—Ru alloy is preferred.

[0047] The thickness of the protective film is not particularly limited as long as it can function as a protective film. From the viewpoint of maintaining the reflectivity of EUV light reflected by the multilayer reflective film, the thickness of the protective film is preferably 1 to 10 nm, more preferably 1.5 to 6 nm, and even more preferably 2 to 5 nm. It is also preferable that the material of the protective film is Ru metal alone, a Ru alloy, Rh metal alone, or a Rh alloy, and that the thickness of the protective film is within the above-mentioned preferred thickness range.

[0048] The protective film may be a film consisting of a single layer, or a multilayer film consisting of multiple layers. When the protective film is a multilayer film, each layer constituting the multilayer film is preferably made of the above-mentioned preferred material. Furthermore, when the protective film is a multilayer film, it is also preferable that the total thickness of the multilayer film is within the above-mentioned preferred range.

[0049] The protective film can be formed by a known film formation method such as magnetron sputtering, ion beam sputtering, etc. When forming a Rh film by magnetron sputtering, it is preferable to use a Rh target as the target and Ar gas as the sputtering gas.

[0050] [Absorber Film] The absorber film of the reflective mask blank is required to have a high contrast between the EUV light reflected by the multilayer reflective film and the EUV light at the absorber film when the absorber film is patterned. The patterned absorber film (absorber film pattern) may function as a binary mask by absorbing EUV light, or may function as a phase shift mask that reflects EUV light and interferes with the EUV light from the multilayer reflective film to generate contrast.

[0051] When the absorber film pattern is used as a binary mask, the absorber film must absorb EUV light and have low reflectance for EUV light. Specifically, when EUV light is irradiated onto the surface of the absorber film, the maximum reflectance of EUV light at a wavelength of around 13.5 nm is preferably 2% or less. The absorber film may contain one or more metals selected from the group consisting of Ta, Ti, Sn, and Cr, as well as one or more components selected from the group consisting of O, N, B, Hf, and H. Among these, it is preferable to contain N or B. By containing N or B, the crystalline state of the absorber film can be made amorphous or microcrystalline. The crystalline state of the absorber film is preferably amorphous. This improves the smoothness and flatness of the absorber film. Furthermore, improving the smoothness and flatness of the absorber film reduces the edge roughness of the absorber film pattern, thereby improving the dimensional accuracy of the absorber film pattern. When the absorber film pattern is used as a binary mask, the film thickness of the absorber film is preferably 40 to 70 nm, more preferably 50 to 65 nm.

[0052] When the absorber film pattern is used as a phase shift mask, the absorber film preferably has a reflectance of 2% or more to EUV light. To obtain a sufficient phase shift effect, the absorber film preferably has a reflectance of 9 to 15%. Using an absorber film as a phase shift mask improves the contrast of the optical image on the wafer and increases the exposure margin. Examples of materials for forming the phase shift mask include Ru metal alone, Ru alloys containing Ru and one or more metals selected from the group consisting of Cr, Au, Pt, Re, Hf, Ta, Ti, and Si, Ta and Nb alloys, oxides containing Ru alloys or TaNb alloys and oxygen, nitrides containing Ru alloys or TaNb alloys and nitrogen, and oxynitrides containing Ru alloys or TaNb alloys and oxygen and nitrogen. Examples of materials for forming the phase shift film include Ir metal alone, and Ir alloys containing Ir and one or more metals selected from the group consisting of Ta, Cr, W, Re, and Si. When the absorber film pattern is used as a phase shift mask, the thickness of the absorber film is preferably 30 to 60 nm, more preferably 35 to 55 nm.

[0053] The absorber film may be a single-layer film or a multilayer film made up of multiple films. When the absorber film is a single-layer film, the number of steps in manufacturing the mask blank can be reduced, thereby improving production efficiency. When the absorber film is a multilayer film, the layer disposed on the opposite side of the absorber film from the protective film side may be an anti-reflection film used when inspecting the absorber film pattern using inspection light (for example, wavelength 193 to 248 nm).

[0054] The absorber film can be formed by a known film formation method such as magnetron sputtering, ion beam sputtering, etc. For example, when a Ta nitride (TaN) film is formed as the absorber film by magnetron sputtering, a Ta target is used, and Ar gas and N ２ The absorber film can be formed by supplying a gas containing the gas and performing sputtering.

[0055] [Conductive film] The reflective mask blank may have a conductive film on the surface (second main surface) opposite to the first main surface of the substrate. By providing the conductive film, the reflective mask blank can be handled using an electrostatic chuck. The conductive film may include one or more first elements selected from the group consisting of Cr and Ta.

[0056] The conductive film contains one or more first elements selected from the group consisting of Cr and Ta. The conductive film may also contain one or more second elements selected from the group consisting of B, C, N, and O. However, the composition of the conductive film is different from the composition of the protective film, which will be described in detail later. Note that "different compositions" not only refers to cases where the conductive film and the protective film contain different elements, but also refers to cases where the conductive film and the protective film contain two or more of the same elements and the content ratios of the elements are different between the conductive film and the protective film. The conductive film preferably contains either Cr or Ta as the first element, and more preferably contains Cr. The conductive film preferably contains N as the second element. It is also preferable that the conductive film contains Cr as the first element and at least N as the second element. Specific materials constituting the conductive film include, for example, Cr alone, CrN, CrO, CrON, CrB, CrBN, CrC, CrCN, CrOC, Ta alone, TaN, TaO, TaON, TaB, TaBN, TaC, TaCN, TaOC, CrTaO, and CrTaN, among which Cr alone, CrN, TaN, or TaBN is preferred. Note that notation such as "CrON" refers to a material containing Cr, O, and N, and the content ratio of these elements is not limited.

[0057] The conductive film preferably has a low sheet resistance. The sheet resistance of the conductive film is, for example, preferably 200 Ω / □ or less, and more preferably 100 Ω / □ or less. The thickness of the conductive film is preferably 10 to 1,000 nm, and more preferably 100 to 500 nm. The back surface conductive film can be formed using a known film formation method, for example, a sputtering method such as magnetron sputtering or ion beam sputtering, a CVD method, a vacuum deposition method, or an electrolytic plating method.

[0058] [Other Films] The reflective mask blank may have other films. Examples of the other films include a hard mask film. The hard mask film is preferably disposed on the side opposite to the protective film side of the absorber film. As the hard mask film, a material having high resistance to dry etching, such as a Cr-based film or a Si-based film, is preferably used. Examples of the Cr-based film include Cr, and materials containing Cr and one or more elements selected from the group consisting of O, N, C, and H. Specific examples include CrO and CrN. Examples of the Si-based film include Si, and materials containing Si and one or more elements selected from the group consisting of O, N, C, and H. Specific examples include SiO ２ , SiON, SiN, SiO, Si, SiC, SiCO, SiCN, and SiCON. When a hard mask film is formed on the absorber film, dry etching can be performed even if the minimum line width of the absorber film pattern is small. Therefore, it is effective for miniaturizing the absorber film pattern. The hard mask film can be formed using a known film formation method such as magnetron sputtering and ion beam sputtering.

[0059] The absorber film of the reflective mask blank is patterned to form an absorber film pattern, thereby obtaining a reflective mask. In the reflective mask, the openings of the absorber film reflect a larger amount of EUV light, and therefore the reflective mask is suitable for use in exposure to EUV light.

[0060] <Mask Blank Manufacturing Method> The method for forming each film of the mask blank is as described above. When the mask blank is a reflective mask blank, it can be manufactured by sequentially forming each of the above films on the first main surface of a mask blank substrate. A conductive film may also be formed on the second main surface. Specifically, a reflective mask blank manufacturing method may include a method for forming a multilayer reflective film on a mask blank substrate, forming a protective film on the multilayer reflective film, forming an absorber film on the protective film, and forming a hard mask film on the absorber film. The mask blank manufacturing method may also include a film manufacturing procedure other than the above. For example, after forming a multilayer reflective film, the intermediate film may be formed on the multilayer reflective film, and the protective film may be formed on the intermediate film. Here, when the intermediate film is formed on the multilayer reflective film after forming the multilayer reflective film, and the protective film is formed on the intermediate film, it is preferable to perform the formation of the intermediate film continuously without exposure to the atmosphere from the start of formation of the intermediate film to the end of formation of the protective film. It is also preferable to perform the formation of the multilayer reflective film continuously without exposure to the atmosphere from the start of formation of the multilayer reflective film to the end of formation of the protective film.

[0061] The mask blank is not limited to a reflective mask blank, but may also be a transmissive mask blank. For example, in a method for manufacturing a transmissive mask blank, a binary mask blank is manufactured by forming a light-shielding film as an absorbing film on the main surface of a mask blank substrate, or a halftone phase shift mask blank is manufactured by forming a semi-transmissive film as an absorbing film, or a halftone phase shift mask blank is manufactured by sequentially forming a semi-transmissive film as an absorbing film and a light-shielding film. A resist film may be formed on these thin films for pattern formation. Here, examples of the light-shielding film generally include Cr films, Cr-based material films selectively containing oxygen, nitrogen, carbon, and fluorine in Cr, laminated films thereof, MoSi films, MoSi-based material films selectively containing oxygen, nitrogen, and carbon in MoSi, laminated films thereof, etc. The phase shift film may be a SiO 2 film having only a phase shift function. ２In addition to the above films, examples of the film include metal silicide oxide films, metal silicide nitride films, metal silicide oxynitride films, metal silicide oxycarbide films, metal silicide oxynitride carbide films (metals: transition metals such as Mo, Ti, W, and Ta), and half-tone films such as CrO films, CrF films, and SiON films, which have a phase shift function and a light transmittance adjusting function (light reducing function).

[0062] 10(a) to 10(d) are schematic cross-sectional views showing, in order of steps, an example of a manufacturing process for a transfer mask using a mask blank substrate according to an embodiment of the present invention. In Fig. 10(a) to 10(d), the same components as those in the mask blank 10 shown in Fig. 1 are denoted by the same reference numerals, and detailed descriptions thereof will be omitted.

[0063] FIG. 10( a) shows a state in which a resist pattern 40 is formed on a mask blank 10 having, in this order, a conductive film 19, a substrate 12, a multilayer reflective film 14, a protective film 16, and an absorber film 18. The resist pattern 40 can be formed using a known method, for example, by applying a resist to the absorber film 18 of the mask blank 10, exposing and developing the resist to form the resist pattern 40. The resist pattern 40 corresponds to a transfer pattern of a resist film formed on a semiconductor substrate using a transfer mask. Then, using the resist pattern 40 in FIG. 10( a) as a mask, the absorber film 18 is etched and patterned, and the resist pattern 40 is removed to form the absorber film pattern 18pt shown in FIG. 10( b), thereby obtaining a laminate having the absorber film pattern 18pt. The absorber film pattern 18pt is a transfer pattern. Next, as shown in Fig. 10(c), a resist pattern 42 corresponding to the frame of the exposed region is formed on the laminate of Fig. 10(b), and dry etching is performed using the resist pattern 42 of Fig. 10(c) as a mask. Dry etching is performed until it reaches the substrate 12. After dry etching, the resist pattern 42 is removed, and a transfer mask 44 shown in Fig. 10(d) is obtained.

[0064] Examples of dry etching used to form the absorber film pattern 18pt include dry etching using a Cl-based gas and dry etching using an F-based gas. The resist pattern 40 and the resist pattern 42 can be removed by known methods, such as removal with a cleaning solution. Examples of cleaning solutions include sulfuric acid-hydrogen peroxide solution (SPM), sulfuric acid, ammonia water, ammonia-hydrogen peroxide solution (APM), OH radical cleaning water, and ozone water. If the mask blank 10 has an etching mask film, the etching mask film may be patterned using the resist pattern 40 as a mask, and dry etching may be performed using the pattern of the etching mask film as a mask. If the mask blank 10 has an etching mask film, a step of removing the etching mask film may be performed in the step of obtaining a transfer mask. Furthermore, the etching mask film may also be removed simultaneously in the step of removing the resist pattern 40 or the resist pattern 42 described above. The absorber film pattern 18pt (transfer pattern) obtained by patterning the absorber film 18 of the mask blank 10 is suitable for a transfer mask 44 used for exposure to EUV light.

[0065] <Method for Manufacturing a Semiconductor Device> A method for manufacturing a semiconductor device includes a step of using a transfer mask obtained by the above-described method for manufacturing a transfer mask as, for example, a reflective mask for EUV exposure, and transferring a transfer pattern (not shown) formed by an absorber film pattern on the reflective mask to a resist film (not shown) formed on a semiconductor substrate (not shown) by exposure using an EUV exposure apparatus (not shown). In the method for manufacturing a semiconductor device, a secondary component of a first virtual surface shape of a first main surface (transfer pattern side) of a mask blank substrate used in manufacturing the reflective mask is set in the exposure apparatus as a mask parameter for transfer exposure, and exposure is performed. The main surface of the reflective mask is the surface on the first main surface side of the mask blank substrate. Although the geometric shape of the main surface of the reflective mask is a BOW (parabolic) shape, exposure can be performed so that it is optically equivalent to a mask surface with high flatness by setting the optical parameters of the exposure apparatus. As a result, the dimensional accuracy and positional accuracy of the transfer pattern formed on the semiconductor device are improved, and semiconductor devices with desired characteristics can be manufactured with a high yield.

[0066] The present invention will be specifically described below using examples, but the present invention is not limited to these examples. Of Examples 1 to 4 of the method for manufacturing a mask blank substrate shown below, Examples 1 and 3 are working examples, and Examples 2 and 4 are comparative examples.

[0067] Example 1: TiO ２ -SiO ２A synthetic quartz substrate (152 mm × 152 mm, square outer dimensions, 6.3 mm thickness) was prepared. The PV value of the substrate in the above-described BOW differential shape was determined before the mask blank substrate manufacturing method was performed. To determine the PV value of the BOW differential shape of the substrate before the mask blank substrate manufacturing method was performed, first, the surface shapes of the first and second main surfaces of the substrate were measured using an optical interferometer (UltraFlat 200, manufactured by Tropel). The measurement area was the entire substrate (152 mm × 152 mm). First virtual surface shape data was calculated using the least squares method from the first main surface shape data representing the surface shape of the first main surface obtained by the optical interferometer. Similarly, second virtual surface shape data was calculated using the least squares method from the second main surface shape data representing the surface shape of the second main surface. Next, the second virtual surface shape data was flipped left and right to create inverted second virtual surface shape data. Next, the first virtual surface shape data and the inverted second virtual surface shape data were added together to obtain the BOW differential shape data. The difference between the maximum and minimum values ​​of the BOW differential shape data is the PV value of the BOW differential shape. Therefore, the maximum and minimum values ​​of the BOW differential shape data were identified, and the difference between the maximum and minimum values ​​was calculated. In Example 1, the PV value of the BOW differential shape was 136.86 nm.

[0068] As described above, the first virtual surface shape data and the second virtual surface shape data were acquired, the first measurement step (step S10) and the second measurement step (step S12) were performed, and the average virtual surface shape calculation step (step S14) was performed partway through. Next, the first virtual surface shape data and the second virtual surface shape data were simply averaged (arithmetic averaged) to obtain average virtual surface shape data. The average virtual surface shape data represents the average virtual surface shape. Next, an average virtual surface 37 (see FIG. 5) having an average virtual surface shape 36 (see FIG. 5) was translated in the thickness direction of the substrate. For example, a first processing reference plane 36a was set as shown in FIG. 7 at a point closest to the average virtual surface 37 (see FIG. 5) in the unevenness of the first main surface 15a of the substrate 15 shown in FIG. 7. The region 38 surrounded by the first main surface 15a and the first processing reference plane 36a was the processing area to be removed. The processing amount of the first main surface was calculated based on the first processing reference plane 36a. Furthermore, an average virtual surface 37 (see FIG. 5) having an average virtual surface shape 36 (see FIG. 5) was translated in the thickness direction of the substrate, and the second machining reference plane 36b was set as shown in FIG. 7 at the point closest to the average virtual surface 37 (see FIG. 5) in the unevenness of the second main surface 15b of the substrate 15 shown in FIG. 7. A region 39 surrounded by the second main surface 15b and the second machining reference plane 36b is the machining region to be removed. The amount of machining of the second main surface was calculated based on the second machining reference plane 36b.

[0069] Next, local processing was performed on the first main surface 15a of the substrate 15 shown in Figure 7 to remove a region 38 based on the processing amount of the first main surface 15a, thereby planarizing the surface. Local processing was performed on the second main surface 15b of the substrate 15 shown in Figure 7 to remove a region 39 based on the processing amount of the second main surface 15b, thereby planarizing the surface. The local processing was performed using a gas cluster ion beam etching method. The substrate after local processing was subjected to finish polishing. The finish polishing was performed using a double-sided polishing device so that the polishing amount was 0.2 µm. The following polishing pad and polishing slurry were used. Polishing pad: ultra-soft polisher Polishing slurry: colloidal silica slurry (pH 1-2) with an average particle size of 20 nm, pH adjusted with nitric acid

[0070] For the substrate after local processing, the PV value in the bow differential shape was determined as described above. In Example 1, the PV value in the bow differential shape of the substrate after local processing was 53.66 nm. For the substrate after finish polishing, the PV value in the bow differential shape was determined as described above. In Example 1, the PV value in the bow differential shape of the substrate after finish polishing was 9.12 nm. Furthermore, the above-mentioned index value was determined for the substrate after finish polishing. The index value was 23.87 nm. When calculating the index value, the first principal surface shape data and second virtual surface shape data of the substrate after finish polishing had already been acquired. Next, the second virtual surface shape data was flipped left and right to create inverted second virtual surface shape data. Next, the first principal surface shape data and the inverted second virtual surface shape data were added together to obtain index bow shape data. Since the difference between the maximum and minimum values ​​of the index bow shape data is the above-mentioned index value, the maximum and minimum values ​​of the index bow shape data were identified, and the difference between the maximum and minimum values ​​was determined. This difference is the index value mentioned above.

[0071] Example 2 In Example 2, a substrate having a PV value of 136.51 nm in the BOW differential shape of the substrate before the mask blank substrate manufacturing method was performed was used. Example 2 differs from Example 1 in the method of setting the first processing reference plane 36a and the second processing reference plane 36b. In Example 2, first virtual surface shape data and second virtual surface shape data were acquired when calculating the PV value in the BOW differential shape of the substrate before the mask blank substrate manufacturing method was performed as described above. In Example 2, the first virtual surface represented by the first virtual surface shape data was defined as the first processing reference plane 36a (see FIG. 7), and the second virtual surface represented by the second virtual surface shape data was defined as the second processing reference plane 36b (see FIG. 7). In Example 2, the area surrounded by the first main surface of the substrate and the first virtual surface represented by the first virtual surface shape data (first processing reference plane 36a) was defined as the processing area, and the processing amount of the first main surface was calculated. Furthermore, the area surrounded by the second main surface of the substrate and the second virtual surface (second processing reference surface 36b) represented by the second virtual surface shape data was defined as the processing area, and the processing amount of the second main surface was calculated. In Example 2, local processing was performed using the local processing device of Example 1 based on the processing amount of the first main surface and the processing amount of the second main surface described above. Furthermore, after the local processing process, a finish polishing process was performed. In Example 2, the PV value in the BOW differential shape of the substrate after local processing was 125.72 nm. The PV value in the BOW differential shape of the substrate after finish polishing was 64.79 nm. Furthermore, the above-mentioned index value of the substrate after finish polishing was 79.79 nm.

[0072] <Example 3> In Example 3, a mask blank substrate manufacturing method was performed in the same manner as in Example 1, except that a substrate was used in which the PV value in the bow differential shape of the substrate before the mask blank substrate manufacturing method was 54.05 nm. In Example 3, the PV value in the bow differential shape of the substrate after local processing was 53.84 nm. The PV value in the bow differential shape of the substrate after finish polishing was 10.27 nm. The above-mentioned index value of the substrate after finish polishing was 27.09 nm. <Example 4> In Example 4, a mask blank substrate manufacturing method was performed in the same manner as in Example 2, except that a substrate was used in which the PV value in the bow differential shape of the substrate before the mask blank substrate manufacturing method was 42.29 nm. In Example 4, the PV value in the bow differential shape of the substrate after local processing was 115.59 nm. The PV value in the bow differential shape of the substrate after finish polishing was 54.66 nm. The above-mentioned index value of the substrate after finish polishing was 66.99 nm.

[0073]

[0074] The results of Examples 1 to 4 are summarized in Table 1 above. As shown in Table 1, among Examples 1 to 4, Examples 1 and 3 had smaller PV values ​​in the bow differential shape of the substrate after local processing compared to Examples 2 and 4. Furthermore, the above-mentioned index value of the substrate after finish polishing was also smaller. As described above, the index value indicates the degree to which the first main surface is flat when the second main surface is in close contact with a horizontal surface. As described above, the mask blank substrate manufacturing method of the present invention was able to manufacture mask blank substrates having flat surfaces in actual use conditions. Furthermore, from Examples 1 and 2 and Examples 3 and 4, it was found that the magnitude of the PV value in the bow differential shape of the substrate before local processing, i.e., before carrying out the mask blank substrate manufacturing method, had little effect on the PV value in the bow differential shape of the substrate after local processing, i.e., after the mask blank substrate was manufactured, and furthermore, the effect on the PV value in the bow differential shape of the substrate after finish polishing was also small.

[0075] The entire contents of the specification, claims, drawings and abstract of Japanese Patent Application No. 2024-134951, filed on August 13, 2024, are incorporated herein by reference as part of the disclosure of the present invention.

[0076] REFERENCE SIGNS LIST 10 Mask blank 12 Substrate 13 Mask blank substrate 14 Multilayer reflective film 15 Substrate 15a First main surface 15b Second main surface 15c Side surface 16 Protective film 18 Absorber film 18pt Absorber film pattern 19 Conductive film 20 Manufacturing system 22 Measurement unit 24 Calculation unit 26 Processing unit 28 Finishing polishing unit 29 Display unit 30, 32 Surface shape 33 First virtual surface shape 34 Average virtual surface shape 35 Second virtual surface shape 36 Average virtual surface shape 36a First processing reference plane 36b Second processing reference plane 37 Average virtual surface 38, 39 Area 40, 42 Resist pattern 44 Transfer mask Hp Horizontal plane

Claims

1. A method for manufacturing a mask blank substrate comprising a substrate having a first main surface and a second main surface opposite to the first main surface, the method comprising: a first measurement step of measuring the surface shape of the first main surface at least in an effective area; a second measurement step of measuring the surface shape of the second main surface at least in the effective area; an average virtual surface shape calculation step of calculating an average virtual surface shape having a second-order component for the substrate as the highest-order component at least in the effective area based on the surface shapes measured in the first measurement step and the second measurement step; and a local processing step of locally processing the first main surface and the second main surface based on the average virtual surface shape.

2. A method for manufacturing a mask blank substrate as described in claim 1, wherein the average virtual surface shape calculation process is a process of obtaining first virtual surface shape data having the second-order component as the highest-order component from the surface shape of the first main surface obtained in the first measurement process, obtaining second virtual surface shape data having the second-order component as the highest-order component from the surface shape of the second main surface obtained in the second measurement process, and averaging the first virtual surface shape data and the second virtual surface shape data to obtain the average virtual surface shape.

3. A method for manufacturing a mask blank substrate as described in claim 1, wherein the average virtual surface shape calculation step is a step of obtaining first main surface shape data from the surface shape of the first main surface obtained in the first measurement step, obtaining second main surface shape data from the surface shape of the second main surface obtained in the second measurement step, averaging the first main surface shape data and the second main surface shape data to obtain average surface shape data, and obtaining the average virtual surface shape having the second-order component as the highest-order component from the average surface shape data.

4. The method for manufacturing a mask blank substrate according to claim 1, wherein the method for calculating the average virtual surface shape in the virtual surface shape calculation step is the least squares method.

5. A method for manufacturing a mask blank substrate according to any one of claims 1 to 4, wherein after finish polishing following the local processing step, the PV value in a BOW differential shape represented by the difference between a first main surface virtual surface shape having the second-order component of the first main surface of the substrate as its highest-order component and a second main surface virtual surface shape having the second-order component of the second main surface of the substrate as its highest-order component is 50 nm or less.

6. The method for manufacturing a mask blank substrate according to any one of claims 1 to 4, wherein the local processing step comprises: a step of translating an average virtual surface having the average virtual surface shape in the thickness direction of the substrate, setting a first processing reference plane at a point in the unevenness of the first main surface of the substrate that is closest to the average virtual surface, and performing local processing to remove an area surrounded by the first main surface and the first processing reference plane; and a step of translating an average virtual surface having the average virtual surface shape in the thickness direction of the substrate, setting a second processing reference plane at a point in the unevenness of the second main surface of the substrate that is closest to the average virtual surface, and performing local processing to remove an area surrounded by the second main surface and the second processing reference plane.

7. The method for manufacturing a mask blank substrate according to claim 1, wherein the local processing is performed using an ion beam etching method.

8. A method for manufacturing a mask blank substrate according to any one of claims 1 to 4, wherein the measurement areas in the first measurement step and the second measurement step coincide when viewed from a direction perpendicular to the first main surface of the substrate.

9. A method for manufacturing a substrate with a multilayer reflective film, comprising a step of forming a multilayer reflective film on the first main surface of the mask blank substrate obtained by the method for manufacturing a mask blank substrate according to any one of claims 1 to 4.

10. A method for producing a mask blank, comprising the step of forming an absorbing film on the multilayer reflective film of a multilayer reflective film coated substrate obtained by the method for producing a multilayer reflective film coated substrate according to claim 9.

11. A method for manufacturing a transfer mask, comprising patterning the absorbing film of the mask blank obtained by the method for manufacturing a mask blank according to claim 10 to form a transfer pattern.

12. A method for manufacturing a semiconductor device, comprising a step of exposing and transferring a transfer pattern onto a resist film on a semiconductor substrate using a transfer mask obtained by the method for manufacturing a transfer mask according to claim 11.

13. A mask blank substrate comprising a substrate having a first main surface and a second main surface opposite to the first main surface, wherein the PV value in a BOW differential shape expressed by the difference between a first main surface virtual surface shape having a second-order component of the first main surface of the substrate as its maximum order component and a second main surface virtual surface shape having a second-order component of the second main surface of the substrate as its maximum order component is 50 nm or less after finish polishing.

Images