Method for evaluating fiducial marks formed on EUV mask blanks

Abstract

To provide a method for evaluating an EUV mask blank that can easily evaluate processing accuracy (for example, depth) of a reference mark.SOLUTION: The present invention relates to a method for evaluating a reference mark formed on an EUV mask blank, the method including steps of: imaging the reference mark formed on the EUV mask blank to obtain a reference mark image; obtaining a reference mark contrast from the obtained reference mark image, the reference mark contrast being contrast between the reference mark and a background level; and evaluating processing accuracy of the reference mark with the obtained reference mark contrast.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a method for evaluating a fiducial mark formed on a reflective mask blank (EUV mask blank) that reflects EUV light and is used in the manufacture of semiconductor devices (semiconductor devices) and the like. [Background technology]

[0002] Semiconductor devices are manufactured using photolithography, a technology in which exposure light is irradiated onto a pattern transfer mask, such as a photomask on which a circuit pattern is drawn, and the circuit pattern formed on the mask is transferred onto a semiconductor substrate via a reduction optical system. As circuit patterns in semiconductor devices become finer, the wavelength of the exposure light has become mainstream, at 193 nm, using argon fluoride (ArF) excimer laser light, and by adopting a process called multi-patterning, which combines exposure and processing processes multiple times, it is possible to form patterns with dimensions significantly smaller than the exposure wavelength. However, as the continuous miniaturization of device patterns has necessitated the formation of even finer patterns, EUV lithography technology has been developed, which uses EUV light, which has an even shorter wavelength than ArF excimer laser light, as exposure light. More specifically, EUV light has a wavelength of around 13.5 nm. Since conventional transmission-type projection optics and masks cannot be used with this EUV light, reflective optical elements are used. For this reason, a reflective mask has been proposed as a mask for pattern transfer.

[0003] A reflective mask has a multilayer reflective film that reflects EUV light formed on a substrate, and an absorber film that absorbs EUV light formed in a pattern on top of the multilayer reflective film. On the other hand, the state before the absorber film is patterned (including the state where a resist layer is formed) is called a reflective mask blank, and this is used as the material for reflective masks. Hereinafter, a reflective mask blank that reflects EUV light will also be referred to as an EUV mask blank. The basic structure of an EUV mask blank is comprised of a multilayer reflective film formed on a low-thermal expansion substrate that reflects EUV light, and an absorber film formed on top of that that absorbs EUV light. The multilayer reflective film is typically a Mo / Si multilayer reflective film, which ensures EUV light reflectivity by alternately stacking molybdenum (Mo) and silicon (Si) films. A protective film is then formed to protect the multilayer reflective film. Meanwhile, the absorber film is typically made of a material primarily composed of tantalum (Ta) or chromium (Cr), which have relatively large extinction coefficients for EUV light.

[0004] One of the challenges facing EUV lithography technology is the lack of information on defect locations in EUV mask blanks. When applying EUV lithography, even a height anomaly of around 1 nm on the surface of the multilayer reflective film of a reflective mask causes a phase shift in the reflected EUV light, resulting in dimensional changes or poor resolution in the transferred pattern when the absorber film pattern is transferred onto a wafer. Height anomalies in the mask blank that cause such phase changes are called phase defects.

[0005] Patent Document 1 discloses a technology in which a fiducial mark is formed on an absorber film, the positions of phase defects in a multilayer reflective film are captured as the positions of irregularities on the surface of the absorber film above it, and the positions are converted into defect position information relative to the fiducial mark, and then drawing data for patterning the absorber film is corrected based on the defect position information, thereby patterning the absorber film while avoiding the positions of the phase defects. However, some phase defects are structural disturbances in the multilayer reflective film that reduce reflectivity and are not easily manifested as irregularities on the surface of the multilayer reflective film, making it difficult to accurately obtain position information for all phase defects in absorber film inspection.

[0006] For this reason, for example, Patent Document 2 proposes forming a fiducial mark on a multilayer reflective film and using this as a reference position to identify the position of the defect. Furthermore, since phase defects depend not only on the surface irregularities of the multilayer reflective film but also on the interior of the reflective multilayer film or the surface irregularities of the low-thermal expansion substrate, inspection methods that use general laser light as inspection light are unable to adequately detect defects. Therefore, it is considered appropriate to use a so-called at-wavelength inspection method, which detects phase defects using inspection light with the same wavelength as the EUV light used to expose the reflective mask. As an example of this method, for example, Patent Document 3 discloses a method using a dark-field inspection image. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] International Publication No. 2014 / 129527 [Patent Document 2] Japanese Patent Application Publication No. 2017-227936 [Patent Document 3] Japanese Patent Application Laid-Open No. 2003-114200 Summary of the Invention [Problem to be solved by the invention]

[0008] The fiducial marks on EUV mask blanks serve as indicators for accurately defining defect positions, so the processing accuracy of these fiducial marks is an important quality parameter. For example, if the fiducial mark cannot be detected by an inspection system because its position is significantly different from the design position, it can be determined that the fiducial mark has been processed incorrectly. However, even if the fiducial mark is correctly detected by the inspection machine, the fiducial mark may be manufactured shallower than designed, resulting in variations in the detection sensitivity of such defective products depending on the inspection machine. If such defective products are released, the fiducial mark may not be detected by the customer's process equipment, resulting in quality defects. Therefore, it is necessary to evaluate the processing accuracy of the fiducial mark. One evaluation method is, for example, measurement using an AFM (atomic force microscope), but AFM measurement is a destructive test and cannot measure the substrate itself used to manufacture the mask.

[0009] The present invention has been made in view of the above-mentioned problems, and aims to provide a method for evaluating an EUV mask blank that can easily evaluate the processing accuracy (for example, depth) of a fiducial mark. [Means for solving the problem]

[0010] In order to achieve the above object, the present invention provides a method for evaluating a fiducial mark formed on an EUV mask blank, comprising the steps of: capturing an image of a reference mark formed on the EUV mask blank to obtain a reference mark image; From the acquired reference mark image, a reference mark contrast, which is a contrast between the reference mark and a background level, is acquired; The present invention provides a method for evaluating a fiducial mark formed on an EUV mask blank, characterized in that the processing accuracy of the fiducial mark is evaluated based on the acquired fiducial mark contrast.

[0011] The "reference mark contrast" referred to here refers to the difference in light intensity between the reference mark portion and the background level portion of an image obtained by shining light on the reference mark of an EUV mask blank and reflecting the light within a specified range of the EUV mask blank. In other words, if the intensity of the reflected light from the reference mark portion is Imark and the intensity of the reflected light from the background level portion is Ibgl, the reference mark contrast Δ is expressed as follows: Δ=(Imark-Ibgl) / (Imark+Ibgl) is defined as:

[0012] As a result of intensive research, the inventors have found that the fiducial mark contrast Δ (hereinafter also referred to simply as contrast Δ) correlates with the fiducial mark processing accuracy (e.g., depth), and that the fiducial mark processing accuracy can be evaluated by measuring the contrast Δ. The present invention is a method for evaluating EUV mask blanks using such a new index, which allows for the non-destructive and simple evaluation of the fiducial mark processing accuracy. Furthermore, the evaluation method of the present invention can be carried out simultaneously with, for example, a defect inspection, thereby enabling full inspection without increasing the evaluation process time. Furthermore, the evaluation method of the present invention makes it possible to eliminate products with abnormalities in the processing accuracy of the reference marks, thereby providing high-quality products.

[0013] Here, when evaluating the processing accuracy of the reference mark, a reference contrast acquisition step of preparing a reference EUV mask blank in which the depth of a reference mark is varied in advance, and acquiring the reference mark contrast of the reference EUV mask blank as a reference contrast; an evaluation object contrast acquisition step of preparing an EUV mask blank to be evaluated that is different from the reference EUV mask blank, and acquiring a reference mark contrast of the evaluation object EUV mask blank as an evaluation object contrast using the same optical system as in the reference contrast acquisition step; A pass / fail judgment step can be performed in which the pass / fail of the reference mark of the EUV mask blank to be evaluated is judged by comparing the contrast to be evaluated with the reference contrast.

[0014] In this way, the depth processing accuracy of the fiducial marks can be evaluated more accurately, and their pass / fail (normal / abnormal) can be determined, allowing for the selection of non-defective EUV mask blanks from among those being evaluated. This makes it possible to more reliably provide EUV mask blanks with stable quality.

[0015] Furthermore, when evaluating the processing accuracy of the reference mark, a reference contrast acquisition step of preparing a reference EUV mask blank in which the depth of a reference mark is varied in advance, and acquiring the reference mark contrast of the reference EUV mask blank as a reference contrast; a correlation data calculation step of acquiring a correlation equation between the reference contrast and the depth of the reference mark from the reference contrast and the depth of the reference mark in the reference EUV mask blank; an evaluation object contrast acquisition step of preparing an EUV mask blank to be evaluated that is different from the reference EUV mask blank, and acquiring a reference mark contrast of the evaluation object EUV mask blank as an evaluation object contrast using the same optical system as in the reference contrast acquisition step; A pass / fail judgment step can be performed to judge the pass / fail of the reference mark of the EUV mask blank to be evaluated by evaluating the depth of the reference mark of the EUV mask blank to be evaluated from the contrast to be evaluated and the correlation equation obtained in the correlation data calculation step.

[0016] In this way, the fiducial mark depth processing accuracy can be evaluated more accurately, and its pass / fail (normal / abnormal) can be determined, allowing for the selection of non-defective EUV mask blanks from among those being evaluated. This makes it possible to more reliably provide EUV mask blanks with stable quality.

[0017] Furthermore, when capturing an image of a reference mark formed on the EUV mask blank to acquire the reference mark image, Schwarzschild optics allows imaging in dark field inspection.

[0018] In this way, when imaging the reference mark, a dark-field inspection image using a Schwarzschild optical system is suitable. [Effects of the Invention]

[0019] The evaluation method of the present invention allows for non-destructive and simple evaluation of the processing accuracy (particularly the depth) of fiducial marks formed on EUV mask blanks, which can eliminate defective fiducial marks and contribute to stable production in customers' mask manufacturing processes. [Brief explanation of the drawings]

[0020] [Figure 1] An example of an evaluation flow of the present invention using a reference mark contrast in EUV mask blank manufacturing is shown. [Figure 2] FIG. 10 is an explanatory diagram showing an example of a reference mark. [Figure 3] 10 is a graph showing the relationship between the measured depth values ​​of the reference marks #1 to #6 and the reference contrast ΔREF. BEST MODE FOR CARRYING OUT THE INVENTION

[0021] Hereinafter, embodiments of the present invention will be described with reference to the drawings, but the present invention is not limited thereto. (First embodiment) First, a case where the fiducial mark contrast is applied as an index for evaluating the processing accuracy of the fiducial marks of an EUV mask blank will be described. The evaluation method of the present invention may include capturing an image of a reference mark formed on an EUV mask blank to obtain a reference mark image, obtaining a reference mark contrast (contrast between the reference mark and a background level) from the obtained reference mark image, and evaluating the processing accuracy of the reference mark on the EUV mask blank using the obtained reference mark contrast. The above-mentioned formula is used to obtain the reference mark contrast (also simply referred to as contrast). In the following, the processing accuracy of the fiducial mark will be described using the processing accuracy of the fiducial mark depth as an example, but the present invention is not limited to this, and any processing accuracy that can be evaluated using the fiducial mark contrast obtained from a fiducial mark image as described below will suffice. Note that the fiducial mark depth can be measured using an AFM or the like.

[0022] It is also possible to perform the above steps only on the EUV mask blank to be evaluated and evaluate the processing accuracy of the fiducial mark depth. However, as will be explained below, in addition to the EUV mask blank to be evaluated, it is also possible to prepare a reference EUV mask blank to use as reference data by changing the contrast conditions by varying the fiducial mark depth conditions, perform the above steps on both of them, and compare the contrasts obtained from each to perform evaluation. In this case, the processing accuracy of the fiducial mark depth can be evaluated more accurately.

[0023] The following describes the process steps in more detail. <Reference contrast acquisition process> First, prepare a reference EUV mask blank. As will be described later, it is possible to prepare one that has the same fiducial mark design as the EUV mask blank (product) to be evaluated, but with a different depth. The EUV mask blank to be actually evaluated has various films on a transparent substrate (glass substrate for photomasks), and the film configuration can be, for example, a Mo / Si multilayer reflective film formed by alternately laminating molybdenum (Mo) and silicon (Si) films to ensure the reflectivity of EUV light, and further a ruthenium (Ru) film formed as a protective film to protect the multilayer reflective film. There are no restrictions on the film formation method, but for example, sputtering is used.

[0024] Next, fiducial marks with different depths are formed on the prepared EUV mask blank, with the same mark design as the product. The fiducial marks can be processed using a variety of methods, such as FIB (focused ion beam) and lithography processes, and the depth conditions can be varied by changing the processing conditions. In this way, a reference EUV mask blank is prepared in which the depths of the fiducial marks vary. The number of depth conditions may be one or more, but it is particularly preferable to have a plurality of conditions. The upper limit of the number is not particularly limited, but can be, for example, 5 to 10. A single reference EUV mask blank may be prepared in which a plurality of fiducial marks of different depths are formed, or a plurality of similar reference EUV mask blanks may be prepared in which a fiducial mark of a different depth is formed on each of the reference EUV mask blanks.

[0025] Next, the fiducial marks of the reference EUV mask blank are imaged to obtain a fiducial mark image. Any device capable of capturing images so that the fiducial mark contrast can be subsequently obtained from the captured image may be used, and for example, a defect inspection device capable of capturing defect images may be used. It is particularly preferable to capture dark-field inspection images using a Schwarzschild optical system, but this is not particularly limited. It is preferable that the imaging conditions are such that there are no defects, such as foreign matter, dents, or pinholes, within the image field including the fiducial mark, since this allows for more accurate evaluation. The wavelength used for imaging is not particularly limited as long as it is possible to obtain a fiducial mark image and fiducial mark contrast, and is preferably 600 nm or less. There is no particular restriction on the lower limit, and for example, EUV light (approximately 1 nm to 100 nm, especially around 13.5 nm) may be used.

[0026] Then, the contrast of each reference mark is obtained from the acquired reference mark images at different depths, and this is set as the reference contrast. It should be noted that, as necessary, an image processing algorithm such as differential processing may be used to further emphasize differences in light intensity, etc., for the image.

[0027] <Evaluation target contrast acquisition process> Next, an EUV mask blank to be evaluated is prepared separately from the reference EUV mask blank used in the reference contrast acquisition step. Then, using the same optical system as in the reference contrast acquisition step, an image of a reference mark formed on the EUV mask blank to be evaluated is acquired in the same manner, and the reference mark contrast is acquired from the reference mark image. This is defined as the contrast to be evaluated.

[0028] <Good / bad judgment process (contrast comparison process)> Next, the processing accuracy of the fiducial mark formed on the EUV mask blank to be evaluated is evaluated by comparing the acquired contrast to be evaluated with the reference contrast, and the quality of the fiducial mark is judged. The standard for the comparison evaluation here is not particularly limited, but for example, the comparison may be made with the acquired reference contrast itself, or a standard set based on the acquired reference contrast may be used. For example, a target contrast may be set with reference to the reference contrast, and a standard within ±10% of the target contrast may be set as the normal value (good product). Alternatively, a device related to the detection of fiducial marks may be used to evaluate reference contrasts obtained by evaluating fiducial marks of different depths on the reference EUV mask blank, and a minimum detectable contrast for that device may be evaluated and set as the normal value (good product) if the minimum detectable contrast is equal to or greater than the reference contrast.

[0029] Using such a standard, the quality of the fiducial marks formed on the EUV mask blank can be determined. With such a standard, fiducial marks with abnormal depth processing accuracy can be more reliably distinguished from those with normal accuracy, and the quality of the evaluation target can be determined with even greater accuracy. Therefore, the quality of the EUV mask blank can be more reliably stabilized. Naturally, the standard is not limited to within ±10% as described above, and a numerical range (threshold) that serves as a standard can be set according to a desired evaluation standard.

[0030] The evaluation method of the present invention as described above makes it possible to evaluate the processing accuracy, such as the depth of a fiducial mark formed on an EUV mask blank, in a non-destructive and simple manner, unlike conventional destructive testing that uses, for example, an AFM. Furthermore, by acquiring a fiducial mark image during defect inspection, it is possible to efficiently inspect all products. By performing this simultaneously with defect inspection, it is possible to perform full inspection without increasing the evaluation process time. In addition, since it is possible to eliminate products with abnormal fiducial mark processing accuracy, it becomes possible to provide high-quality EUV mask blanks. It is possible to provide high-quality products stably and easily.

[0031] Furthermore, when the evaluation method of the present invention is incorporated into defect inspection, the present invention can be carried out according to the following flow. FIG. 1 shows an example of the process of the evaluation flow of the present invention using the reference mark contrast in EUV mask blank manufacturing. The evaluation device can be a defect inspection device capable of acquiring a reference mark image. First, a program is constructed to automate the series of steps of acquiring a reference mark image, acquiring contrast, and determining whether the image is good or bad (contrast comparison).

[0032] Next, as a preliminary test, a reference contrast acquisition process is performed on a reference EUV mask blank, including substrate loading, fiducial mark image acquisition, and fiducial mark contrast acquisition. That is, a reference EUV mask blank with fiducial marks of different depths is loaded into the system, fiducial mark images are acquired, and the reference contrast ΔREF is calculated. From the reference contrast ΔREF, a target depth contrast ΔSTD is set, and finally, a threshold value is set in the evaluation system that determines a product to be acceptable if it is within ΔSTD ±10%.

[0033] After completing the above settings, the EUV mask blank to be evaluated is inspected. In the reference mark image acquisition process, a reference mark image is acquired from the EUV mask blank to be evaluated that has been loaded into the equipment, and the reference mark contrast (contrast to be evaluated) is automatically measured from the acquired image. Then, in the pass / fail judgment process, the contrast of the inspected product (contrast to be evaluated) is compared with a preset standard value (contrast ΔSTD set from reference contrast ΔREF), and only if it is judged to be within ±10%, the process proceeds to the next step. If an EUV mask blank to be evaluated is judged to be abnormal, the inspection is stopped and an alarm is generated.

[0034] This inspection flow makes it possible to quickly detect abnormalities in processing accuracy, such as the depth of fiducial marks formed on EUV mask blanks, which were previously difficult to determine whether they were good or bad, and enables rapid process improvements.

[0035] (Second embodiment) Next, a case where the correlation between the fiducial mark contrast and the fiducial mark depth is used as an index for evaluating the processing accuracy of the fiducial mark formed on the EUV mask blank will be described. In the first embodiment, the fiducial mark contrast itself was used as the index, but in this embodiment, the fiducial mark depth, which has a strong correlation with the fiducial mark contrast and is obtained by conversion from the fiducial mark contrast, is used as the index.

[0036] <Reference contrast acquisition process> A reference EUV mask blank is prepared in advance, on which fiducial marks with varying depth conditions are formed. Next, the fiducial marks are imaged, and the fiducial mark contrasts (reference contrasts) under each depth condition are obtained from the fiducial mark images. The depth of the reference mark can be measured by AFM or the like. The acquisition of the reference mark image and the reference contrast can be performed in the same manner as in the first embodiment.

[0037] <Correlation data calculation process> From the reference contrast and the depth of the reference mark in the reference EUV mask blank obtained as described above, a correlation equation between the reference contrast and the depth of the reference mark is obtained.

[0038] <Evaluation target contrast acquisition process> Next, an EUV mask blank to be evaluated, separate from the reference EUV mask blank, is prepared, and a reference mark image is obtained by capturing an image of the reference mark on the EUV mask blank to be evaluated using the same optical system as in the reference contrast acquisition step, and the reference mark contrast is obtained from the acquired reference mark image (contrast to be evaluated). The acquisition of the reference mark image and the evaluation target contrast can be performed in the same manner as in the first embodiment.

[0039] <Good / bad judgment process> The depth of the fiducial mark of the EUV mask blank to be evaluated is evaluated from the contrast to be evaluated using the correlation equation between the reference contrast and the fiducial mark depth obtained in the correlation data calculation step. In this way, the depth processing accuracy of the fiducial mark formed on the EUV mask blank can be evaluated non-destructively and simply, and the quality of the fiducial mark can be determined.

[0040] In the second embodiment, similar to the first embodiment, the various effects described above can be achieved. In evaluating the depth of the fiducial marks on the EUV mask blank to be evaluated, a standard that the depth is within a predetermined value % of the target depth (this predetermined value is not limited, but is particularly within ±10%) can be set as a normal value (good product). Alternatively, fiducial marks of different depths on the reference EUV mask blank described above can be evaluated using an apparatus that uses fiducial marks, and the minimum detectable depth for that apparatus can be evaluated and determined, and a depth equal to or greater than the minimum detectable depth (or deeper than the minimum detectable depth) can be set as a normal value (good product). [Example]

[0041] The present invention will be explained in more detail below by showing examples and comparative examples, but the present invention is not limited to these examples. Example 1 Among the methods for evaluating the fiducial marks in the present invention, an evaluation method using the fiducial mark contrast was carried out. First, a molybdenum (Mo) film and a silicon (Si) film were alternately laminated on a square glass substrate measuring 6 inches (approximately 15 cm) on a side using a sputtering method using a molybdenum target and a silicon target as the target and argon as the sputtering gas, to form a Mo / Si multilayer reflective film that ensures EUV light reflectivity. Furthermore, a ruthenium (Ru) film was formed as a protective film to protect the multilayer reflective film, and a reference EUV mask blank was produced. Next, fiducial marks #1 to #6 with different depths were formed on one reference EUV mask blank by varying the processing conditions using an FIB. The actual depths of these fiducial marks were measured using an AFM. Table 1 shows the actual measured depths of fiducial marks #1 to #6. Based on the actual measured values, fiducial marks with depths of 30 nm to 80 nm were formed.

[0042] [Table 1]

[0043] Additionally, an image of the fiducial mark was acquired using ABICS (E120) manufactured by Lasertec Corp. From the acquired image, the fiducial mark contrast (contrast ΔA to ΔH) calculated from the aforementioned formula, which is the difference in light intensity from the background level, was calculated for sides A to H of one fiducial mark in the design shown in Figure 2, and the average value of these ΔA to ΔH was calculated as the reference contrast ΔREF. Table 2 shows the results of ΔA to ΔH and ΔREF for fiducial marks #1 to #6, and FIG. 3 shows a graph plotting the measured depth values ​​and reference contrast ΔREF for fiducial marks #1 to #6.

[0044] [Table 2]

[0045] From this result, it was confirmed that the reference contrast ΔREF has a correlation with the depth of the reference mark. Here, the target contrast ΔSTD was set to ΔSTD=1.19.

[0046] Next, for EUV mask blanks LOT (a), (b), and (c) formed under the same film formation conditions as the reference EUV mask blank, four fiducial marks, MARK1, MARK2, MARK3, and MARK4, were created for each LOT using the FIB under processing conditions that targeted ΔSTD=1.19 (target depth of 50 nm), with the same design as #1 to #6 shown in FIG. 2, and these were used as EUV mask blanks to be evaluated. Images of the fiducial marks were acquired under the same illumination conditions using the ABICS manufactured by Lasertec Corporation. Then, from the acquired images, the fiducial mark contrasts (contrasts ΔA to ΔH) were calculated from the sides A to H shown in FIG. 2 and the background level, and the average value of these was used to calculate the evaluation target contrast Δ AVE was calculated. Table 3 shows the Δ of MARK1 to 4 AVE Shows.

[0047] [Table 3]

[0048] Based on these results, the standard for determining a non-defective product was set as the contrast to be evaluated being within ±10% (1.07 to 1.31) of the target contrast ΔSTD. Based on this set criterion, MARK1 of LOT(c) was determined to be abnormal, and only LOT(a)(b) was determined to be non-defective.

[0049] Example 2 Among the methods for evaluating the fiducial marks in the present invention, an evaluation method was carried out using a depth value converted from a correlation equation between the fiducial mark contrast and the fiducial mark depth. The following correlation equation was calculated from FIG. 3, which is a plot graph of the depth and reference contrast ΔREF of the reference marks #1 to #6 evaluated in Example 1. Correlation formula: Reference mark depth = 149.93 x ΔREF - 127.48

[0050] Next, using the above correlation formula, the contrasts to be evaluated for MARK1 to MARK4 of LOTs (a) to (c) fabricated in Example 1 with a target depth of 50 nm, ΔAVE to depth D AVE Table 4 shows the depth D of MARK1 to MARK4. AVE Shows.

[0051] [Table 4]

[0052] Based on these results, the standard for determining whether a product is acceptable was set at a target value of 50 nm ±10% (45 nm to 55 nm). Based on this set criterion, MARK1 of LOT(c) was determined to be abnormal, and only LOT(a)(b) was determined to be non-defective.

[0053] (verification) As described above, the target depth of the fiducial marks in Examples 1 and 2 is 50 nm (range of depth processing accuracy in non-defective products: 45 to 55 nm). Therefore, when the depths of the reference marks in LOT(a), (b), and (c) were actually measured by AFM, the actual measured values ​​were the same as the results shown in Table 4. That is, similar to the judgment results in Examples 1 and 2, only MARK1 in LOT(c) was abnormal, the others were normal, and LOT(a) and (b) were non-defective. This demonstrates that the evaluation method of the present invention is effective.

[0054] The present invention is not limited to the above-described embodiments. The above-described embodiments are merely examples, and any configuration that is substantially identical to the technical idea described in the claims of the present invention and that provides similar effects is included within the technical scope of the present invention. [Explanation of symbols]

[0055] A~F...Edges that make up the reference mark.

Claims

1. A method for evaluating a fiducial mark formed on an EUV mask blank, comprising: capturing an image of a reference mark formed on the EUV mask blank to obtain a reference mark image; From the acquired reference mark image, a reference mark contrast, which is a contrast between the reference mark and a background level, is acquired; When evaluating the processing accuracy of the reference mark using the acquired reference mark contrast, When evaluating the processing accuracy of the reference mark, a reference contrast acquisition step of preparing a reference EUV mask blank in which the depth of a reference mark has been varied in advance, and acquiring the reference mark contrast of the reference EUV mask blank as a reference contrast; a correlation data calculation step of acquiring a correlation equation between the reference contrast and the depth of the reference mark from the reference contrast and the depth of the reference mark in the reference EUV mask blank; an evaluation object contrast acquisition step of preparing an EUV mask blank to be evaluated that is different from the reference EUV mask blank, and acquiring a reference mark contrast of the evaluation object EUV mask blank as an evaluation object contrast using the same optical system as in the reference contrast acquisition step; a pass / fail determination step of determining whether the reference mark of the EUV mask blank to be evaluated is pass / fail by evaluating the depth of the reference mark of the EUV mask blank to be evaluated from the contrast to be evaluated and the correlation equation obtained in the correlation data calculation step.

2. When capturing an image of a reference mark formed on the EUV mask blank to acquire the reference mark image, 2. The method for evaluating a fiducial mark formed on an EUV mask blank according to claim 1, wherein the image is captured as a dark-field inspection image using a Schwarzschild optical system.

Images