Surface shape measurement method, mask blank substrate inspection method, and mask blank manufacturing method

Abstract

To provide a technique for improving measurement accuracy of a surface shape.SOLUTION: A method for measuring a surface shape has: measuring a surface shape of a first substrate by a measuring instrument; detecting an error between measurement data of the surface shape of the first substrate and reference data of the surface shape of the first substrate; measuring the surface shape of one or more second substrates prepared separately from the first substrate by the measuring instrument; and correcting the measurement data of the surface shape of the second substrate on the basis of the error. The method for measuring a surface shape has: detecting an abnormal point caused by a defect of the first substrate created after acquiring the reference data, in measurement data corrected on the basis of the error; and smoothing a contour curve showing the abnormal point and the surface shape in the vicinity of the point.SELECTED DRAWING: Figure 6

Description

[Technical Field]

[0001] The present disclosure relates to a method for measuring a surface shape, a method for inspecting a mask blank substrate, and a method for manufacturing a mask blank. [Background technology]

[0002] In recent years, with the miniaturization of semiconductor devices, EUV lithography (EUVL), an exposure technology using extreme ultraviolet (EUV), has been developed. EUV includes soft X-rays and vacuum ultraviolet light, and specifically refers to light with a wavelength of approximately 0.2 nm to 100 nm. At present, EUV with a wavelength of approximately 13.5 nm is mainly being considered.

[0003] A reflective mask is used in EUVL. A reflective mask has, in that order, a substrate such as a glass substrate, a multilayer reflective film that reflects EUV, and an absorbing film that absorbs EUV. An aperture pattern is formed in the absorbing film. In EUVL, the aperture pattern in the absorbing film is transferred to a target substrate such as a semiconductor substrate. Transferring includes reducing and transferring.

[0004] A method for manufacturing a mask blank for EUVL includes a step of measuring the surface shape of a substrate, and a step of locally processing the substrate to improve the flatness of the substrate by referring to the measurement results of the surface shape of the substrate (see, for example, Patent Document 1). The surface shape of the substrate is measured using a non-contact measuring device such as a laser interference type. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Patent Publication No. 2021-181398 Summary of the Invention [Problem to be solved by the invention]

[0006] The surface shape measurement method includes measuring the surface shape of a first substrate with a measuring instrument and detecting an error between the measurement data of the surface shape of the first substrate and reference data of the surface shape of the first substrate. The first substrate is a reference substrate for correction. The reference data of the surface shape of the first substrate is measured in advance with another measuring instrument and stored in the measuring instrument as a true value.

[0007] The error between the reference data and the measurement data is caused by the structure of the measuring instrument, for example, by wavefront aberration. The first substrate is stored inside the measuring instrument and is periodically subjected to measurement of its surface shape. Error detection is then performed periodically. This is because the error may change over time as the condition of the measuring instrument changes over time.

[0008] The surface shape measurement method includes measuring the surface shape of one or more second substrates prepared separately from the first substrate using a measuring device, and correcting the measurement data of the surface shape of the second substrate based on the error. By correcting the measurement data based on the error, errors caused by the structure of the measuring device can be reduced.

[0009] However, after obtaining reference data for the surface shape of the first substrate, defects may occur on the first substrate. The defects may be convex or concave. A convex defect may be, for example, an attachment such as a particle. A concave defect may be, for example, a scratch. When the first substrate has a defect, the surface shape of the first substrate may be measured, and an error between the measurement data and the reference data may be detected.

[0010] In this case, the detected error will include a component caused by the defect on the first substrate. If the measurement data of the surface shape of the second substrate is then corrected based on the error, an abnormal point caused by the defect on the first substrate will appear in the corrected measurement data of the second substrate. The abnormal point is an illusion that does not actually exist. This reduces the measurement accuracy of the surface shape.

[0011] One aspect of the present disclosure provides a technique for improving the accuracy of measuring surface shapes. [Means for solving the problem]

[0012] A surface shape measurement method according to one aspect of the present disclosure includes measuring the surface shape of a first substrate with a measuring device, detecting an error between the measurement data of the surface shape of the first substrate and reference data for the surface shape of the first substrate, measuring the surface shapes of one or more second substrates prepared separately from the first substrate with the measuring device, and correcting the measurement data of the surface shapes of the second substrates based on the error.The surface shape measurement method also includes detecting, in the measurement data corrected based on the error, an abnormal point caused by a defect in the first substrate that was added after the reference data was acquired, and smoothing a contour curve indicating the surface shape of the abnormal point and its vicinity. [Effects of the Invention]

[0013] According to one aspect of the present disclosure, by smoothing the surface shape of an abnormal point and its vicinity, it is possible to blur abnormal points that do not actually exist but appear to exist, thereby improving the measurement accuracy of the surface shape. [Brief explanation of the drawings]

[0014] [Figure 1] FIG. 1 is a flowchart showing a method for manufacturing a mask blank according to one embodiment. [Figure 2] FIG. 2 is a cross-sectional view showing an example of a mask blank substrate. [Figure 3] FIG. 3 is a plan view of the mask blank substrate of FIG. [Figure 4] FIG. 4 is a cross-sectional view showing an example of a mask blank. [Figure 5] FIG. 5 is a cross-sectional view showing an example of a mask. [Figure 6] FIG. 6 is a flowchart showing a method for measuring a surface shape according to an embodiment. [Figure 7]Figure 7(A) is a diagram showing measurement data of the surface shape of the second substrate, Figure 7(B) is a diagram showing measurement data when the second substrate of Figure 7(A) is rotated 90°, Figure 7(C) is a diagram showing measurement data when the second substrate of Figure 7(A) is rotated 180°, and Figure 7(D) is a diagram showing measurement data when the second substrate of Figure 7(A) is rotated 270°. [Figure 8] Figure 8(A) is a diagram showing the measurement data of the surface shape of the first second substrate, Figure 8(B) is a diagram showing the measurement data of the surface shape of the second second substrate, and Figure 8(C) is a diagram showing the measurement data of the surface shape of the third second substrate. [Figure 9] Figure 9(A) shows measurement data corrected based on error, Figure 9(B) shows a portion of the measurement data in Figure 9(A), Figure 9(C) shows smoothed data obtained by smoothing the measurement data in Figure 9(B), and Figure 9(D) is a diagram created by replacing a portion of the measurement data in Figure 9(A) with the replacement data in Figure 9(C). [Figure 10] FIG. 10 is a diagram illustrating an example of the gain of a Gaussian filter. DETAILED DESCRIPTION OF THE INVENTION

[0015] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings. In each drawing, the same or corresponding components are denoted by the same reference numerals, and their description may be omitted. In the specification, the symbol "to" indicating a range of values ​​means that the values ​​before and after it are included as the lower and upper limits.

[0016] A method for manufacturing a mask blank according to one embodiment will be described with reference to Fig. 1. The mask blank is not particularly limited, but is, for example, a reflective mask blank for EUVL. As shown in Fig. 1, the method for manufacturing a mask blank includes, for example, steps S101 to S107. Before step S101, a substrate 210, for example, as shown in Figs. 2 and 3, is prepared.

[0017] The substrate 210 includes a first main surface 211 and a second main surface 212 facing opposite to the first main surface 211. The first main surface 211 is rectangular. In this specification, a rectangular shape includes a shape with chamfered corners. A rectangle also includes a square. The second main surface 212 faces opposite to the first main surface 211. Like the first main surface 211, the second main surface 212 is also rectangular.

[0018] The substrate 210 also includes four end surfaces 213, four first chamfered surfaces 214, and four second chamfered surfaces 215. The end surfaces 213 are perpendicular to the first main surface 211 and the second main surface 212. The first chamfered surfaces 214 are formed at the boundaries between the first main surface 211 and the end surfaces 213. The second chamfered surfaces 215 are formed at the boundaries between the second main surface 212 and the end surfaces 213. In this embodiment, the first chamfered surfaces 214 and the second chamfered surfaces 215 are so-called C-chamfered surfaces, but may also be R-chamfered surfaces.

[0019] The substrate 210 is, for example, a glass substrate. The glass of the substrate 210 is preferably quartz glass containing TiO2. Compared to common soda-lime glass, quartz glass has a smaller linear expansion coefficient and undergoes less dimensional change due to temperature changes. The quartz glass may contain 80% to 95% by mass of SiO2 and 4% to 17% by mass of TiO2. When the TiO2 content is 4% to 17% by mass, the linear expansion coefficient is approximately zero near room temperature, and there is almost no dimensional change near room temperature. The quartz glass may contain a third component or impurity other than SiO2 and TiO2.

[0020] The size of the substrate 210 in plan view is, for example, 152 mm in length and 152 mm in width. The length and width may be greater than or equal to 152 mm.

[0021] The substrate 210 has a central region 211A and a peripheral region 211B on its first main surface 211. The central region 211A is a square region excluding a rectangular frame-shaped peripheral region 211B surrounding the central region 211A, and is a region that is processed to a desired flatness in steps S101 to S104 and is a quality assurance region. The quality assurance region has a size of, for example, 142 mm in length and 142 mm in width. The length and width dimensions may be 142 mm or more. The four sides of the central region 211A are parallel to the four end faces 213. The center of the central region 211A coincides with the center of the first main surface 211.

[0022] Although not shown, the second main surface 212 of the substrate 210 also has a central region and a peripheral region, similar to the first main surface 211. The central region of the second main surface 212 is a square region, similar to the central region of the first main surface 211, and is a region that is processed to a desired flatness in steps S101 to S104 of FIG. 1, and is a quality assurance region. The quality assurance region has a size of, for example, 142 mm in length and 142 mm in width. The length and width dimensions may be 142 mm or more.

[0023] Step S101 includes polishing the first main surface 211 and the second main surface 212 of the substrate 210. In this embodiment, the first main surface 211 and the second main surface 212 are polished simultaneously using a double-sided polisher (not shown), but they may also be polished sequentially using a single-sided polisher (not shown). In step S101, the substrate 210 is polished while supplying polishing slurry between the polishing pad and the substrate 210.

[0024] Examples of the polishing pad include a urethane-based polishing pad, a nonwoven fabric-based polishing pad, and a suede-based polishing pad. The polishing slurry contains an abrasive and a dispersion medium. The abrasive is, for example, cerium oxide particles. The dispersion medium is, for example, water or an organic solvent. The first main surface 211 and the second main surface 212 may be polished multiple times using abrasives of different materials or particle sizes.

[0025] The abrasive used in step S101 is not limited to cerium oxide particles, and may be, for example, silicon oxide particles, aluminum oxide particles, zirconium oxide particles, titanium oxide particles, diamond particles, or silicon carbide particles.

[0026] Step S102 includes measuring the surface shapes of the first main surface 211 and the second main surface 212 of the substrate 210. To measure the surface shapes, for example, a non-contact measuring machine is used to prevent the surface from being scratched. The measuring machine measures the surface shapes of the central region 211A of the first main surface 211 and the central region of the second main surface 212.

[0027] Step S103 includes locally processing the first main surface 211 and the second main surface 212 of the substrate 210 to improve flatness with reference to the measurement results of step S102. The first main surface 211 and the second main surface 212 are locally processed in that order. The order in which they are processed may be any order, and is not particularly limited.

[0028] For the local processing, for example, at least one method selected from a gas cluster ion beam (GCIB) method, a plasma chemical vaporization machining (PCVM) method, a polishing method using a magnetic fluid, and polishing using a rotary polishing tool is used.

[0029] Step S104 includes performing finish polishing of the first main surface 211 and the second main surface 212 of the substrate 210. In this embodiment, the first main surface 211 and the second main surface 212 are polished simultaneously using a double-sided polisher (not shown), but they may also be polished sequentially using a single-sided polisher (not shown). In step S104, the substrate 210 is polished while a polishing slurry is supplied between the polishing pad and the substrate 210. The polishing slurry contains an abrasive. The abrasive is, for example, colloidal silica particles.

[0030] Step S105 includes forming a conductive film 240 shown in FIG. 4 in a central region of the second main surface 212 of the substrate 210. The conductive film 240 is used to attach a mask 201 (see FIG. 5), which will be described later, to an electrostatic chuck of an exposure tool. The conductive film 240 is made of, for example, chromium nitride (CrN). The conductive film 240 is formed by, for example, sputtering.

[0031] Step S106 includes forming a multilayer reflective film 220 shown in FIG. 4 in the central region 211A of the first main surface 211 of the substrate 210. The multilayer reflective film 220 reflects EUV. The multilayer reflective film 220 is formed by alternately stacking high-refractive index layers and low-refractive index layers, for example. The high-refractive index layers are made of, for example, silicon (Si), and the low-refractive index layers are made of, for example, molybdenum (Mo). The multilayer reflective film 220 can be formed by a sputtering method such as ion beam sputtering or magnetron sputtering.

[0032] Step S107 includes forming an absorbing film 230 shown in FIG. 4 on the multilayer reflective film 220 formed in step S106. The absorbing film 230 absorbs EUV. The absorbing film 230 may be a phase shift film and may shift the phase of EUV. The absorbing film 230 is formed of, for example, a single metal, alloy, nitride, oxide, oxynitride, or the like containing at least one element selected from tantalum (Ta), chromium (Cr), and palladium (Pd). The absorbing film 230 is formed, for example, by sputtering.

[0033] In this embodiment, steps S106 and S107 are performed after step S105, but may be performed before step S105.

[0034] 4 is obtained by the above steps S101 to S107. The mask blank 200 has, in this order, a conductive film 240, a substrate 210, a reflective multilayer film 220, and an absorbing film 230. Note that the mask blank 200 may include other films in addition to the conductive film 240, the substrate 210, the reflective multilayer film 220, and the absorbing film 230.

[0035] For example, the mask blank 200 may further include a low-reflection film. The low-reflection film is formed on the absorbing film 230. Then, an opening pattern 231 is formed in both the low-reflection film and the absorbing film 230. The low-reflection film is used for inspecting the opening pattern 231, and has lower reflectivity to inspection light than the absorbing film 230. The low-reflection film is made of, for example, TaON or TaO. The low-reflection film is formed, for example, by sputtering.

[0036] The mask blank 200 may further include a protective film. The protective film is formed between the multilayer reflective film 220 and the absorbing film 230. The protective film protects the multilayer reflective film 220 so that the multilayer reflective film 220 is not etched when the absorbing film 230 is etched to form the opening pattern 231 in the absorbing film 230. The protective film is formed of, for example, Ru, Si, or TiO2. The protective film is formed by, for example, sputtering.

[0037] 5 is obtained by forming an opening pattern 231 in an absorbing film 230. Photolithography and etching are used to form the opening pattern 231. Therefore, the resist film used to form the opening pattern 231 may be included in the mask blank 200.

[0038] An example of a method for measuring a surface shape will be described with reference to Fig. 6. The method for measuring a surface shape includes measuring the surface shape of a first substrate with a measuring instrument (step S201) and detecting an error between the measurement data of the surface shape of the first substrate and reference data of the surface shape of the first substrate (step S202). The first substrate is a reference substrate for correction. The reference data of the surface shape of the first substrate is measured in advance with another measuring instrument and stored in the measuring instrument as a true value. The reference data of the surface shape of the first substrate is measured by a certification body such as NIST.

[0039] The error between the reference data and the measurement data is caused by the structure of the measuring instrument, for example, by wavefront aberration. The first substrate is stored inside the measuring instrument and is periodically subjected to measurement of its surface shape. Error detection is then performed periodically. This is because the error may change over time as the condition of the measuring instrument changes over time.

[0040] The measuring instrument is not particularly limited, but for example, a Fizeau interferometer. A Fizeau interferometer measures the surface shape of a substrate by utilizing the interference between a reference light beam reflected from a reference plane and a test light beam that passes through the reference plane and is reflected from the substrate surface. The Fizeau interferometer is not particularly limited, but for example, an oblique incidence interferometer. In an oblique incidence interferometer, the reference plane is one surface of a triangular prism.

[0041] The surface shape measurement method includes measuring the surface shape of one or more second substrates prepared separately from the first substrate using a measuring device (step S203), and correcting the measurement data of the surface shape of the second substrate based on the error (step S204). By correcting the measurement data of the surface shape of the second substrate based on the error, errors caused by the structure of the measuring device can be reduced.

[0042] Corrections to reduce errors due to the structure of the measuring machine are common, and therefore a detailed description thereof will be omitted. Note that step S204 may further include corrections to reduce errors due to gravitational deformation of the second substrate. This correction is also common, and therefore a detailed description thereof will be omitted.

[0043] The second substrate is a substrate for measurement. The second substrate is not particularly limited, but may be, for example, a mask blank substrate 210. The timing for measuring the surface shape of the substrate 210 is not particularly limited, but may be, for example, step S102 in Fig. 1. The measurement data measured in step S102 is corrected based on the error and then used in step S103 in Fig. 1.

[0044] The most recent error detected in the past is used to correct the measurement data of the surface shape. This is because errors can change over time. Typically, the surface shapes of many second substrates are measured between the nth (n is a natural number greater than or equal to 1) error detection and the n+1th error detection. The error detected the nth time is used to correct the measurement data measured between the nth error detection and the n+1th error detection.

[0045] However, after obtaining reference data for the surface shape of the first substrate, defects may occur on the first substrate. The defects may be convex or concave. A convex defect may be, for example, an attachment such as a particle. A concave defect may be, for example, a scratch. When the first substrate has a defect, the surface shape of the first substrate may be measured, and an error between the measurement data and the reference data may be detected.

[0046] In this case, the detected error includes a component caused by the defect on the first substrate. If the measurement data of the surface shape of the second substrate is then corrected based on the error, an abnormal point caused by the defect on the first substrate will appear in the corrected measurement data. The abnormal point that appears in the corrected measurement data of the second substrate and the defect on the first substrate have an inverted shape.

[0047] If the defect on the first substrate is convex, the abnormal point that appears in the corrected measurement data of the second substrate will be concave. On the other hand, if the defect on the first substrate is concave, the abnormal point that appears in the corrected measurement data of the second substrate will be convex. In either case, the abnormal point is an illusion that does not actually exist.

[0048] Therefore, the surface shape measurement method includes detecting abnormal points due to defects on the first substrate that occurred after the reference data was acquired (step S205), and smoothing the contour curves that indicate the surface shape of the abnormal points and their vicinity (step S206). Smoothing can blur the abnormal points, improving the measurement accuracy of the surface shape.

[0049] Step S205 may include comparing multiple measurement data obtained by measuring the surface shape of the same second substrate while changing the orientation of the second substrate, as shown in Figures 7(A) to 7(D). In Figures 7(A) to 7(D), the height difference on the substrate surface is represented by a grayscale, with lower brightness indicating lower height. The same applies to Figures 8 and 9. The height difference on the substrate surface may be represented by color. Alternatively, the height difference on the substrate surface may be represented by contour lines.

[0050] The abnormal point P caused by the defect in the first substrate does not actually exist on the second substrate. Therefore, even if the second substrate is rotated and the surface shape of the second substrate is measured, the position of the abnormal point P does not rotate. The abnormal point P is observed as a fixed point. Note that Figures 7(A) to 7(D) are image data showing the surface shape of the same second substrate measured at different rotation angles, so if the image is rotated, the image data will be the same except for the abnormal point P.

[0051] Step S205 may include comparing the measurement data of the surface shapes of two or more second substrates (three in FIG. 8) as shown in FIGS. 8(A) to 8(C). An abnormal point P caused by a defect in the first substrate does not actually exist on the second substrate. Therefore, even if the surface shape of a different second substrate is measured, the position of the abnormal point P does not change. The abnormal point P is observed as a fixed point. Note that, because FIGS. 8(A) to 8(C) are image data showing the surface shapes of different second substrates, the image data will not be the same whether they are rotated or not.

[0052] It is preferable to use image data showing elevation differences as the measurement data to be compared in step S205. By displaying the image data on a display, the similarities and differences between the image data can be visually confirmed. Multiple image data may be displayed side by side at the same time, or may be displayed in order in the same area without being displayed simultaneously. In the latter case, the position of the abnormal point P, which is a fixed point, is easy to see. The number of image data to be compared should be two or more, but is preferably three to ten.

[0053] Step S206 involves smoothing the contour curve that indicates the surface shape of the abnormal point P and its vicinity. First, measurement data D2 (see FIG. 9(B)) of a partial region is cut out from measurement data D1 (see FIG. 9(A)) of the quality assurance region that has been corrected based on the error. The cut-out region includes the abnormal point P at its center. The cut-out region is determined appropriately depending on the smoothing method, but is, for example, a square region with a side length of 10 mm to 50 mm.

[0054] Next, the contour curve representing the surface shape of the extracted region is smoothed to create smoothed data D3 (see FIG. 9(C)) in which the abnormal points P are blurred. After that, replacement data D4 is created by extracting a portion of the smoothed data D3. The replacement data D4 is intended to replace a portion of the original measurement data D1.

[0055] The region for creating the replacement data D4 includes the abnormal point P at its center. The region for creating the replacement data D4 is smaller than the region for creating the smoothed data D3. The abnormal point P and its vicinity are smoothed while the irregularities in other regions are preserved. The region for creating the replacement data D4 is, for example, a square region with a side length of 0.5 mm to 3 mm.

[0056] Finally, measurement data D5 (see FIG. 9(D)) with improved measurement accuracy is created by replacing a portion of the original measurement data D1 with replacement data D4. The created measurement data D5 is used, for example, for local processing of the mask blank substrate 210 (step S103 in FIG. 1). This allows the flatness of the substrate 210 to be improved with high accuracy.

[0057] Step S206 may include using a first Gaussian filter. Using the first Gaussian filter includes, for example, extracting wavelength components equal to or greater than a first threshold from a contour curve that indicates the surface shape of the abnormal point and its vicinity. The first threshold is preferably 1.0 mm to 10.0 mm. By extracting wavelength components equal to or greater than the first threshold, wavelength components less than the first threshold can be removed, thereby achieving smoothing.

[0058] Specifically, based on JIS B 0634:2017, a normal distribution curve with the cutoff wavelength (λc) as the first threshold is convolved with the contour curve to obtain a contour curve that extracts low-frequency components (i.e., long-wavelength components). When the first threshold is 5.0 mm, the amplitude transmission ratio of the low-pass filter is as shown by the solid line in Figure 10, and the gain is 0.5 at the cutoff wavelength (λc) of 5.0 mm. Note that no cutoff constant is used in the convolution.

[0059] The method for smoothing the surface shape of the anomalous point P and its vicinity is not particularly limited. For example, a filter with a gain distribution different from a Gaussian distribution may be used. The gain may be the same for all wavelengths, and an averaging filter may be used. Furthermore, after deleting the surface shape of the anomalous point P and its vicinity, the surface shape of the anomalous point P and its vicinity may be reconstructed by interpolation.

[0060] As described above, step S206 includes processing the measurement data D1 corrected based on the error to smooth the contour curve representing the surface shape of the abnormal point P and its vicinity. However, the technology of the present disclosure is not limited to this. Step S206 may also include smoothing the contour curve representing the surface shape of the abnormal point P and its vicinity by correcting the error itself before correcting the measurement data based on the error. Correcting the error includes reducing components of the error that are attributable to defects in the first substrate.

[0061] The surface shape measuring method may include, after step S206, extracting wavelength components equal to or less than a second threshold value from the contour curve representing the surface shape of the second substrate using a second Gaussian filter (not shown). The second threshold value is preferably 5.0 mm to 15.0 mm. The second threshold value may be equal to or greater than the first threshold value. By extracting wavelength components equal to or less than the second threshold value, wavelength components exceeding the second threshold value can be removed.

[0062] Specifically, based on JIS B 0634:2017, a new contour curve is obtained by convolving a normal distribution curve with a cutoff wavelength (λc) as the second threshold with the contour curve, and then subtracting the resulting new contour curve from the original contour curve to obtain a contour curve that extracts high-frequency components (i.e., short-wavelength components).When the second threshold is 5.0 mm, the amplitude transmission ratio of the high-pass filter is as shown by the dashed line in Figure 10, and the gain is 0.5 at the cutoff wavelength (λc) of 5.0 mm.Note that the cutoff constant is set to 0.7 when performing the convolution integral.

[0063] By extracting the short wavelength component using the second Gaussian filter, it is possible to extract important information in the local processing (step S103 in FIG. 1) of the substrate 210. The local processing is processing that removes the short wavelength component.

[0064] The smoothing in step S206 preferably reduces the maximum height difference (PV value) of the components extracted using the second Gaussian filter compared to when smoothing is not performed. The PV value is the maximum height difference when a plane approximated by the least squares method is used as the reference height. The smaller the PV value, the better the flatness.

[0065] As is clear from Table 1 described later, smoothing using the first Gaussian filter can reduce the PV value of the component extracted using the second Gaussian filter.

[0066] As shown in Table 1, the surface profile measurement method may include preparing multiple combinations of a first substrate and a measuring machine, and determining the combination of a first substrate and a measuring machine that has a defect by comparing measurement data of the surface profile of the same second substrate measured using the multiple measuring machines. Here, the measurement data to be compared has been corrected based on errors.

[0067] [Table 1]

[0068] As mentioned above, the errors detected in advance may include components caused by defects on the first substrate. In this case, if the measurement data of the surface shape of the second substrate is corrected based on the errors, abnormal points caused by the defects on the first substrate will appear in the corrected measurement data of the second substrate, degrading the flatness. Flatness is expressed by the so-called PV value. The larger the PV value, the worse the flatness.

[0069] Here, the flatness is preferably the flatness of wavelength components extracted using a second Gaussian filter from the measurement data after error correction, and more preferably the flatness is the flatness of wavelength components extracted using a second Gaussian filter from the wavelength components extracted using a first Gaussian filter from the measurement data after error correction.

[0070] Determining the combination of the first substrate with the defect and the measuring machine may include comparing the flatness of the same second substrate measured using multiple measuring machines. In Table 1, the combination with the worst flatness is set 4. Therefore, the combination of the first substrate with the defect and the measuring machine is determined to be set 4.

[0071] The flatness may be compared by quantifying it as shown in Table 1, or it may be compared as image data. Image data showing height differences may be used as measurement data to compare when determining the combination of the first substrate with defects and the measuring device.

[0072] From the viewpoint of improving reliability, the number of combinations of first substrates and measuring machines to be prepared is preferably 3 or more. Furthermore, from the viewpoint of reducing costs, the number of combinations of first substrates and measuring machines to be prepared is preferably 10 or less.

[0073] The surface profile measurement method, mask blank substrate inspection method, and mask blank manufacturing method according to the present disclosure have been described above, but the present disclosure is not limited to the above-described embodiments. Various changes, modifications, substitutions, additions, deletions, and combinations are possible within the scope of the claims. These naturally fall within the technical scope of the present disclosure. [Explanation of symbols]

[0074] 210 Mask blank substrate (second substrate)

Claims

1. A method for measuring a surface shape, comprising: measuring a surface shape of a first substrate with a measuring device; detecting an error between measurement data of the surface shape of the first substrate and reference data of the surface shape of the first substrate; measuring the surface shapes of one or more second substrates prepared separately from the first substrate with the measuring device; and correcting the measurement data of the surface shape of the second substrate based on the error, detecting an abnormal point caused by a defect on the first substrate that occurred after the reference data was acquired in the measurement data corrected based on the error; smoothing a contour curve representing the surface shape of the abnormal point and its vicinity; A method for measuring a surface shape comprising:

2. The method for measuring a surface shape according to claim 1 , wherein detecting the abnormal point includes comparing a plurality of measurement data obtained by measuring the surface shape of the same second substrate while changing the orientation of the second substrate.

3. The method for measuring a surface shape according to claim 1 , wherein detecting the abnormal point includes comparing measurement data of the surface shapes of two or more of the second substrates.

4. 4. The method for measuring a surface shape according to claim 2, further comprising using image data indicating a height difference as the measurement data to be compared.

5. The method of claim 1 , wherein the smoothing step includes using a first Gaussian filter.

6. using the first Gaussian filter includes extracting wavelength components equal to or greater than a first threshold value from a contour curve indicating a surface shape of the abnormal point and its vicinity; 6. The method for measuring a surface shape according to claim 5, wherein the first threshold value is 1.0 mm to 10.0 mm.

7. After the smoothing, extracting wavelength components equal to or less than a second threshold value from a contour curve representing the surface shape of the second substrate using a second Gaussian filter; 2. The method for measuring a surface shape according to claim 1, wherein the second threshold value is 5.0 mm to 15.0 mm.

8. 8. The surface shape measurement method according to claim 7, wherein the smoothing reduces a maximum peak-to-valley (PV value) of the component extracted using the second Gaussian filter compared to a case where the smoothing is not performed.

9. preparing a plurality of combinations of the first substrate and the measuring device; and determining a combination of the first substrate having the defect and the measuring machine by comparing measurement data of the surface shape of the same second substrate measured using a plurality of the measuring machines.

10. The method for measuring a surface shape according to claim 9 , wherein determining the combination includes comparing flatnesses of the same second substrate measured using a plurality of the measuring machines.

11. 10. The method for measuring a surface shape according to claim 9, further comprising using image data indicating a height difference as the measurement data to be compared.

12. The surface shape measuring method according to claim 9 , wherein the number of combinations of the first substrate and the measuring machine is three or more.

13. 2. The method for measuring a surface shape according to claim 1, wherein the measuring instrument is a Fizeau interferometer.

14. A method for inspecting a mask blank substrate, comprising using the method for measuring a surface shape according to claim 1.

15. A method for manufacturing a mask blank, comprising using the surface shape measurement method according to claim 1.

Images