Evaluation method of crystal defect of silicon single crystal substrate

Abstract

To provide a method for evaluating the number of a crystal defect of a high-aspect ratio existed on a substrate front surface without depending on a type of defects, and provide a method for evaluating defect of a size that is less than an inspection limit size of a front surface inspection device.SOLUTION: An evaluation method for a crystal defect of a silicon single crystal substrate, contains: a first step of measuring the silicon single crystal substrate with a mirror surface polishing with a front surface inspection device, and acquiring position information of a foreign object; a second step of classifying the foreign object into the crystal defect and a foreign object other than the crystal defect, and evaluating a type of crystal defect; a third step of inspecting whether or not an inspection size of the crystal defect is an aerotropic defection that is different depending on a defection direction of the crystal defect, and evaluating a front surface shape of the maximum inspection size; a fourth step of calculating an existence ratio of the front surface shape on a front surface of the silicon single crystal substrate; a fifth step of estimating the number that the aerotropic defection having the front surface shape of the maximum inspection size exists; and a sixth step of estimating the number of the aerotropic defection having the front surface shape that is smaller than the maximum inspection size.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a method for evaluating crystal defects of a silicon single crystal substrate.

Background Art

[0002] Defects present on the surface of a semiconductor substrate (hereinafter sometimes simply referred to as "substrate") cause device failures. Therefore, with the progress of miniaturization due to the sophistication of devices in recent years, reduction and evaluation of defects have become more important.

[0003] Defects present on the substrate surface are mainly detected by a surface inspection apparatus. The surface inspection apparatus inspects the substrate surface by irradiating light onto the substrate surface and detecting scattered light due to foreign matters such as defects. The intensity of the scattered light varies depending on the size, shape, composition, etc. of the foreign matter. For example, the larger the foreign matter, the greater the scattered light intensity.

[0004] In addition, in the surface inspection apparatus, the detected size can be calculated based on the scattered light intensity and the scattered light intensity for each size of standard particles. That is, the larger the scattered light intensity, the larger the detected size. In order to detect foreign matters with a small detected size, it is necessary to perform measurement by the surface inspection apparatus for a long time. Also, in the case of foreign matters with a very small scattered light intensity, it cannot be distinguished from the scattered light caused by the roughness (haze) of the substrate surface, and thus cannot be detected.

[0005] Among the defects present in the substrate, it is known that the shape of the crystal defects varies depending on the crystal growth conditions and the impurity concentration. As a technique for evaluating anisotropic defects, Patent Document 1 discloses a technique for detecting anisotropic defects by irradiating light from a specific direction with respect to the crystal orientation of the substrate.

[0006] Also, as a technique for visualizing crystal defects that are difficult to detect with a surface inspection apparatus, Patent Document 2 discloses a technique for visualizing crystal defects by anisotropic etching.

[0007] In addition, Patent Document 3 discloses a technique for revealing crystal defects by thermal oxidation treatment.

[0008] In addition, Patent Document 4 discloses a technique for revealing crystal defects by epitaxial growth on the substrate surface.

Prior Art Documents

Patent Documents

[0009]

Patent Document 1

Patent Document 2

Patent Document 3

Patent Document 4

Summary of the Invention

Problems to be Solved by the Invention

[0010] Crystal defects having a three-dimensional shape with a high aspect ratio (for example, rod-shaped defects) have different sizes and shapes when exposed on the substrate surface depending on the orientation of the major axis direction with respect to the substrate surface (hereinafter, may be referred to as "defect orientation"). Therefore, depending on the defect orientation, the scattered light intensity becomes small, making it difficult to detect with a surface inspection device. Even in the case of defects with a small size that are difficult to detect with a surface inspection device and are exposed on the substrate surface, if they extend deeply inside the substrate, they can cause device failures. Therefore, there is a need for a method to evaluate the degree to which high aspect ratio crystal defects that are difficult to detect with a surface inspection device exist on the substrate surface.

[0011] The technique described in Patent Document 1 had a problem that it was impossible to detect defects smaller than the detection limit size of the surface inspection device.

[0012] In addition, the technologies described in Patent Documents 2 to 4 are technologies for visualizing specific crystal defects, respectively. Therefore, it is impossible to apply them to the entire crystal defects with a high aspect ratio. In particular, all of these technologies include a heating process, and there is a problem that it is impossible to apply them to thermally unstable defects.

[0013] The present invention has been made in view of the above problems, and an object thereof is to provide a method for evaluating the number of crystal defects with a high aspect ratio present on the surface of a substrate regardless of the type of defect, and a method for evaluating defects having a size smaller than the detection limit size of a surface inspection apparatus.

Means for Solving the Problems

[0014] The present invention has been made to achieve the above object, and includes a first step of measuring a mirror-polished single-crystal silicon substrate with a surface inspection apparatus, detecting foreign matters on the single-crystal silicon substrate, and acquiring position information of the foreign matters; a second step of observing the surface shape of the foreign matters based on the position information of the foreign matters, classifying the foreign matters into crystal defects and foreign matters other than the crystal defects, and evaluating the types based on the shape of the crystal defects; a third step of verifying, for each type of crystal defect based on the shape, whether the detected size of the crystal defect is an anisotropic defect that varies depending on the defect orientation of the crystal defect, and for the crystal defects evaluated as the anisotropic defects, evaluating, for each type of the anisotropic defects, the surface shape having the maximum detected size among the surface shapes on the surface of the single-crystal silicon substrate for each equivalent defect orientation of the anisotropic defects; a fourth step of calculating the abundance ratio of the surface shape on the surface of the single-crystal silicon substrate for each equivalent defect orientation of the anisotropic defects for each type of the anisotropic defects from the three-dimensional shape and orientation of the anisotropic defects; a fifth step of estimating the number of the anisotropic defects having the surface shape with the maximum detected size present on the surface of the single-crystal silicon substrate; and a sixth step of estimating the number of the anisotropic defects having a surface shape smaller than the maximum detected size from the abundance ratio calculated in the fourth step and the number of the anisotropic defects having the surface shape with the maximum detected size present on the surface of the single-crystal silicon substrate estimated in the fifth step. A method for evaluating crystal defects of a single-crystal silicon substrate is provided, characterized by including the above steps.

[0015] According to such a method for evaluating crystal defects of a silicon single crystal substrate, the number of high aspect ratio crystal defects present on the substrate surface can be evaluated regardless of the type of defects, and defects having a size smaller than the detection limit size of the surface inspection apparatus can also be evaluated.

[0016] At this time, the second step can include observation of the surface shape of the foreign matter using a microscope.

[0017] Thereby, small crystal defects that have been a problem in semiconductor substrates for recent state-of-the-art devices can be observed.

[0018] At this time, the second step can include observation of the shape of a cross section of the crystal defect perpendicular to the surface of the silicon single crystal substrate, which intersects the crystal defect, using a transmission electron microscope or a scanning transmission electron microscope.

[0019] Thereby, the three-dimensional shape and orientation of the crystal defect can be evaluated from the shape of the cross section.

[0020] At this time, the third step can be evaluated based on the three-dimensional shape of a known crystal defect.

[0021] Thereby, the three-dimensional shape and orientation of each crystal defect can be evaluated from the surface shape of the crystal defect.

[0022] At this time, the third step can be evaluated based on the three-dimensional shape of the crystal defect evaluated by observation of the shape of the cross section.

[0023] Thereby, the three-dimensional shape and orientation of each crystal defect can be evaluated from the surface shape and cross-sectional shape of the crystal defect.

Advantages of the Invention

[0024] As described above, according to the method for evaluating crystal defects of the silicon single crystal substrate of the present invention, the number of high aspect ratio crystal defects present on the substrate surface can be evaluated regardless of the type of defects, and defects having a size smaller than the detection limit size of the surface inspection apparatus can also be evaluated.

Brief Description of the Drawings

[0025]

Figure 1

Figure 2

[0110] present on the substrate surface.

Figure 3

[0110] present on the substrate surface.

Figure 4

[0011] present on the substrate surface.

Figure 5

[0011] present on the substrate surface.

Figure 6

Figure 7

Figure 8

Figure 9

[0101] present on the substrate surface.

Figure 10

[0101] present on the substrate surface.

Figure 11

[0110] present on the substrate surface.

Figure 12

[0110] present on the substrate surface.

Figure 13

Embodiments for Carrying Out the Invention

[0026] Hereinafter, the present invention will be described in detail, but the present invention is not limited thereto.

[0027] As described above, there has been a demand for a method of evaluating the number of high aspect ratio crystal defects present on the substrate surface regardless of the type of defect, and a method of evaluating defects having a size less than the detection limit size of the surface inspection apparatus.

[0028] As a result of intensive studies on the above problems, the present inventors measured a mirror-polished single-crystalline silicon substrate with a surface inspection apparatus, and detected foreign matter on the single-crystalline silicon substrate and acquired position information of the foreign matter in a first step. In a second step, the surface shape of the foreign matter is observed based on the position information of the foreign matter, the foreign matter is classified into crystal defects and foreign matter other than the crystal defects, and the type based on the shape of the crystal defects is evaluated. For each type based on the shape of the crystal defects, in a third step, it is verified whether the detected size of the crystal defects is an anisotropic defect that varies depending on the defect orientation of the crystal defects. For the crystal defects evaluated as the anisotropic defects, for each type of the anisotropic defects, the surface shape of the maximum detected size among the surface shapes on the surface of the single-crystalline silicon substrate for each equivalent defect orientation of the anisotropic defects is evaluated. For each type of the anisotropic defects, in a fourth step, the abundance ratio of the surface shape on the surface of the single-crystalline silicon substrate for each equivalent defect orientation of the anisotropic defects is calculated from the three-dimensional shape and orientation of the anisotropic defects. In a fifth step, the number of the anisotropic defects having the surface shape of the maximum detected size present on the surface of the single-crystalline silicon substrate is estimated. In a sixth step, the number of the anisotropic defects having a surface shape smaller than the maximum detected size is estimated from the abundance ratio calculated in the fourth step and the number of the anisotropic defects having the surface shape of the maximum detected size present on the surface of the single-crystalline silicon substrate estimated in the fifth step. According to the method for evaluating crystal defects of a single-crystalline silicon substrate, which includes the above steps, the number of high aspect ratio crystal defects present on the substrate surface can be evaluated regardless of the type of the defects, and it has been found that defects having a size smaller than the detection limit size of the surface inspection apparatus can be evaluated, and thus the present invention has been completed.

[0029] [Method for Evaluating Crystal Defects of Single-Crystalline Silicon Substrate] FIG. 1 shows an example of a process flow of a method for evaluating crystal defects of a silicon single crystal substrate according to an embodiment of the present invention. As shown in FIG. 1, the method for evaluating crystal defects of the silicon single crystal substrate of the present invention includes a first step of acquiring position information of foreign substances, a second step of classifying the foreign substances into crystal defects and foreign substances other than crystal defects and evaluating the types of crystal defects, a third step of verifying whether an anisotropic defect in which the detected size of the crystal defect varies depending on the defect orientation of the crystal defect and evaluating the surface shape of the maximum detected size, a fourth step of calculating the abundance ratio of the surface shape on the surface of the silicon single crystal substrate, a fifth step of estimating the number of anisotropic defects having the surface shape of the maximum detected size, and a sixth step of estimating the number of anisotropic defects having a surface shape smaller than the maximum detected size.

[0030] (Step of acquiring position information of foreign substances (first step)) A silicon single crystal substrate on which high aspect ratio crystal defects are formed is used as a sample. The surface of the substrate is made mirror-like by mirror polishing.

[0031] After mirror polishing the substrate, cleaning may be performed. By performing cleaning, particles attached to the substrate surface can be removed. Removing the particles reduces the number of foreign substances observed in the second step, so the time required for the second step can be shortened.

[0032] Next, the substrate is measured with a surface inspection apparatus to detect foreign substances on the substrate and acquire the position information (hereinafter also referred to as "coordinates") of the foreign substances. The surface inspection apparatus can be, for example, Surfscan SP7 manufactured by KLA-Tencor.

[0033] At the time of the measurement, the detected size of the foreign substance may be acquired in accordance with the coordinates of the foreign substance. Acquisition of the detected size of the foreign substance by the surface inspection apparatus can be performed simultaneously with the acquisition of the coordinates of the foreign substance. By acquiring the detected size, it is possible to specify the surface shape having the maximum detected size based on the actual detected size in the fifth step.

[0034] Some surface inspection devices are equipped with a plurality of detectors or detection methods with different detection angles and polarization characteristics of scattered light. The coordinates and detection size of foreign matter may be obtained using a single detector or detection method, or may be obtained using a combination of a plurality of detectors or detection methods with the above-described surface inspection device.

[0035] By using a single detector or detection method, the influence on the detection size due to differences in the detector or detection method can be eliminated. Therefore, when obtaining the detection size, it is desirable to use a single detector or detection method.

[0036] By using a combination of a plurality of detectors or detection methods, foreign matter that cannot be detected by a single detector or detection method can be detected. Therefore, when the actual detection size is not important, it is desirable to use a combination of a plurality of detectors or detection methods.

[0037] (Step of classifying foreign matter into crystal defects and foreign matter other than crystal defects and evaluating the type of crystal defect (second step)) Next, based on the position information (coordinates) of the foreign matter obtained in the first step, the surface shape of the foreign matter (hereinafter also simply referred to as "surface shape") is observed, and the foreign matter is classified into crystal defects and foreign matter other than crystal defects. As foreign matter other than crystal defects, for example, particles, defects due to processing, etc. exist.

[0038] The surface shape of the foreign matter can be observed using various microscopes such as a scanning electron microscope (SEM), a transmission electron microscope (TEM), a scanning transmission electron microscope (STEM), an atomic force microscope (AFM), a laser microscope, etc., but it is particularly desirable to use an electron microscope. By using such a microscope, small crystal defects that are problematic in semiconductor substrates for recent state-of-the-art devices can be observed.

[0039] Among electron microscopes, a review SEM that can automatically perform SEM observation based on the coordinates of foreign objects is particularly suitable for observing foreign objects. By using a review SEM, surface observation can be performed in a shorter time compared to TEM and AFM. The review SEM can be, for example, the eDR7280 manufactured by KLA-Tencor.

[0040] The number of foreign objects to observe the surface shape is desirably all foreign objects at the coordinates detected in the first step, but is not limited thereto. When there are many detected coordinates, it can also be a plurality of randomly extracted foreign objects. The number of foreign objects to be extracted when randomly extracting can be, for example, 300, but is not particularly limited as long as it is a statistically sufficiently effective number for estimating the three-dimensional shape and orientation of high aspect ratio crystal defects existing on the substrate surface and the number by detection size.

[0041] Crystal defects have unique shapes for each type, and the side surfaces of three-dimensional crystal defects are planes with equivalent crystal orientations. High aspect ratio crystal defects exist in a single crystal with the long axis direction parallel to one of the side surfaces forming the defect. And crystal defects with defect orientations having equivalent crystal orientations but different directions are exposed on the substrate surface. At this time, the phenomenon that the surface shape of the crystal defects formed on the substrate surface and the relative size on the substrate surface are different due to the difference in crystal orientation is called orientation.

[0042] Next, for the foreign objects classified as crystal defects, the type based on the shape (three-dimensional shape) of the crystal defects (hereinafter, also simply referred to as "type of crystal defects") is evaluated. Here, those having the same composition (including voids) and three-dimensional shape of the crystal defects are regarded as the same type of crystal defects even if the surface shape is different on the substrate surface due to the defect orientation. Here, examples of the three-dimensional shape include a sphere, a regular octahedron, a flat plate shape, a rod shape, a nanoribbon shape, etc., but are not limited thereto.

[0043] When the three-dimensional shape and orientation of a crystal defect are unknown, the cross-sectional shape of the crystal defect (hereinafter, also simply referred to as the "cross-sectional shape") may be observed. The cross-section can be a cross-section perpendicular to the surface of the silicon single crystal substrate intersecting the crystal defect, but it may also be a cross-section intersecting obliquely with respect to the substrate surface.

[0044] In order to observe the cross-section, it is necessary to fabricate a cross-section at the position of the defect. As an apparatus for performing such processing, a focused ion beam (FIB) apparatus or an argon ion milling apparatus can be used. In particular, in order to accurately form a cross-section including the defect, it is desirable to use an FIB apparatus with high processing position accuracy.

[0045] As an apparatus for observing the cross-sectional shape, for example, SEM, TEM, STEM, etc. can be used. In particular, TEM and STEM, which can observe the cross-sectional shape of the entire defect by transmission observation, are suitable.

[0046] By using TEM or STEM, the crystal defect can be observed with higher resolution than SEM. Since TEM and STEM are for transmission observation, they are particularly suitable for observing the cross-sectional shape while maintaining the structure of the crystal defect.

[0047] By observing the cross-sectional shape, even when the three-dimensional shape and orientation of the crystal defect are unknown, the three-dimensional shape and orientation can be investigated. Therefore, when the three-dimensional shape and orientation of the crystal defect are unknown, it is preferable to observe the cross-sectional shape.

[0048] The observation of the cross-sectional shape for investigating the three-dimensional shape and orientation can be performed for all crystal defects whose surface shape has been observed. However, since the observation of the cross-sectional shape takes a longer time compared to the observation of the surface shape, it is desirable to select a part of the crystal defects. Here, for the observation of crystal defects with a high aspect ratio, it is desirable that the orientation can be confirmed by including at least one of all the existing defect orientations for each defect. By confirming the orientation of the crystal defects to be observed in this way, in the third step, the three-dimensional shape of the crystal defect can be investigated in more detail.

[0049] (Step (the third step) of verifying whether the defect is an anisotropic defect in which the detection size of the crystal defect varies depending on the defect orientation of the crystal defect and evaluating the surface shape of the maximum detection size) Next, for each type (type of crystal defect) based on the shape of the crystal defect evaluated in the second step, verify whether the detection size of the crystal defect is an anisotropic defect that varies depending on the defect orientation of the crystal defect. For the crystal defects evaluated as anisotropic defects, for each type of anisotropic defect, evaluate the surface shape of the maximum detection size among the surface shapes on the silicon single crystal substrate surface for each equivalent defect orientation of the anisotropic defect.

[0050] First, for each type of crystal defect, verify whether the detection size of the crystal defect varies depending on the defect orientation of the crystal defect.

[0051] When the three-dimensional shape of the crystal defect has a low aspect ratio, the surface shape does not change significantly depending on the defect orientation. Examples of three-dimensional shapes with a low aspect ratio include a sphere and a regular octahedron.

[0052] When the three-dimensional shape of the crystal defect has a high aspect ratio, the surface shape and the relative size on the substrate surface may vary depending on the defect orientation. For crystal defects having a three-dimensional shape with a high aspect ratio, for each type of crystal defect, from the three-dimensional shape and the orientation, obtain the surface shape and the relative size on the substrate surface when exposed on the substrate surface, and verify whether the detection sizes are different. Examples of three-dimensional shapes with a high aspect ratio include a rod shape and a nanoribbon shape.

[0053] In the case of a crystal defect with a known three-dimensional shape and orientation, it is possible to determine the three-dimensional shape and orientation of each crystal defect from the surface shape. However, the aspect ratio of the crystal defect shape varies depending on the growth conditions of the single crystal and the impurity concentration. Therefore, it is desirable to obtain the aspect ratio of the crystal defect shape for each substrate from the observation results of the surface shape. At this time, if the substrates can be regarded as having the same single crystal growth conditions and impurity concentrations, the same aspect ratio can be applied.

[0054] Accordingly, the three-dimensional shape and orientation of each crystal defect can be evaluated from the surface shape of the crystal defect.

[0055] On the other hand, in the case of a crystal defect with an unknown three-dimensional shape and orientation, the three-dimensional shape, orientation, and aspect ratio of each crystal defect are investigated from the surface shape and cross-sectional shape of the crystal defect observed in the second step. The crystal defects for observing the cross-sectional shape can be all of the crystal defects for observing the surface shape, or can be a part thereof. When it is a part, it is necessary to obtain the aspect ratio by observing the cross-sectional shape for at least one defect orientation for each different surface shape.

[0056] Accordingly, the three-dimensional shape and orientation of each crystal defect can be evaluated from the surface shape and cross-sectional shape of the crystal defect.

[0057] The steps up to this point will be described using the examples shown in FIGS. 2 to 5. As an example, a case where each form of defect orientations

[0011] , [0-11],

[0110] ,

[0101] , [1-10], [10-1] of a rod-shaped defect whose long axis direction of the defect is equivalent to the

[0110] direction exists in a (100) plane silicon single crystal substrate will be given as an example.

[0058] At this time, when the defect orientation is

[0011] or [0-11], as shown in FIG. 2, the defect orientation is parallel to the (100) plane, and the entire rod-shaped defect 70 is cut parallel to the longitudinal direction by the (100) plane. Therefore, if the length in the direction perpendicular to the longitudinal direction of the rod-shaped defect 70 is 1 (the same length 1 in both the planar direction and the depth direction of the surface of the substrate 80), and the length in the longitudinal direction is a, the aspect ratio of the surface shape is 1 to a as shown in FIG. 3. Here, a may be any value as long as the three-dimensional shape of the crystal defect is rod-shaped, and can be, for example, 2 or more.

[0059] On the other hand, when the defect orientation is

[0110] ,

[0101] , [1-10] or [10-1], as shown in FIG. 4, the body portion of the rod-shaped defect 70 is cut obliquely at 45° by the (100) plane. Therefore, as shown in FIG. 5, when the length in the direction perpendicular to the plane direction of the substrate 80 surface with respect to the longitudinal direction is set to 1, the aspect ratio of the surface shape is approximately 1:(cos45°) -1 = 1:2^0.5

[0060] As described above, the aspect ratio of the shape of the (100) plane of the rod-shaped defect 70 whose major axis direction of the defect is equivalent to the

[0110] direction is 1:a (a≧2) when the defect orientation is

[0011] or [0-11], and 1:2^0.5 when the defect orientation is

[0110] ,

[0101] , [1-10] or [10-1], and the two are clearly distinguishable. Furthermore, since the surface geometric area of the (100) plane of the rod-shaped defect 70 is larger in the former case, it is determined that the detection size of the rod-shaped defect 70 differs depending on the defect orientation.

[0061] The crystal defect thus evaluated that the detection size of the crystal defect differs depending on the defect orientation of the crystal defect is referred to as an anisotropic defect.

[0062] Next, the surface shape with the maximum detection size among the surface shapes on the silicon single crystal substrate surface for each equivalent defect orientation of the anisotropic defect is evaluated.

[0063] Here, for the surface shape with the maximum detection size, the aspect ratio of the surface shape with a large geometric area on the substrate surface for each equivalent defect orientation of the anisotropic defect is obtained, and the surface shape with the largest aspect ratio can be used. This is because generally, the larger the size of the defect formed on the substrate surface, the stronger the scattering intensity with respect to the incident light.

[0064] For example, in the case of a rod-shaped defect 70 existing in a silicon single crystal substrate 80 having the (100) plane, where the major axis direction of the defect is equivalent to the

[0110] direction, there are surface shapes with an aspect ratio of 1 to a (a ≥ 2) and 1 to √2. When the size of the entire defect is the same, since the geometric surface area is larger for the former, it can be evaluated that the surface shape with an aspect ratio of 1 to a (a ≥ 2) has the largest detection size.

[0065] Instead of the above, the surface shape with the largest detection size can be determined as the one with the largest average value or median value calculated for each surface shape from the measurement results of the surface inspection device.

[0066] By basing on the average value of the actual detection size, a result that conforms to the actual situation can be obtained. Also, by basing on the median value of the actual detection size, more accurate evaluation is possible even when there are defects so large that the detection size is overloaded.

[0067] In addition, when there are multiple types of anisotropic defects on the substrate to be evaluated, the surface shape with the largest detection size is evaluated for each type of anisotropic defect.

[0068] (Step (fourth step) of calculating the abundance ratio of surface shapes on the silicon single crystal substrate surface) Next, for each type of anisotropic defect, the abundance ratio of the surface shapes on the silicon single crystal substrate surface for each equivalent defect orientation of the anisotropic defect is calculated from the three-dimensional shape and orientation of the anisotropic defect.

[0069] Specifically, first, from the three-dimensional shape and aspect ratio of the anisotropic defect obtained in the third step, for each equivalent defect orientation of the anisotropic defect, the ratio of the maximum length in the vertical direction (hereinafter sometimes simply referred to as "depth") that is longer than the length in the minor axis direction of the defect with respect to the substrate surface to the length in the minor axis direction of the anisotropic defect is obtained.

[0070] Since the surface of the substrate is a plane cut out at a certain position of the crystal, when the existence ratios of the respective defect orientations with equivalent crystal orientations are the same in the silicon single crystal, defects with a long defect orientation are likely to be exposed on the surface of the cut-out substrate. The probability of its existence increases in proportion to the depth.

[0071] Here, the existence ratio of the surface shape that a certain type of crystal defect can take on the substrate surface is the product of the ratio of the number of defect orientations that can be taken for each surface shape and the ratio of the depth to the length in the minor axis direction of the crystal defect for each defect orientation. Therefore, the ratio of the number of equivalent crystal orientations of the anisotropic defect is multiplied by the ratio of the depth to the length in the minor axis direction for each equivalent defect orientation of the anisotropic defect obtained above to calculate the existence ratio of the surface shape on the surface of the silicon single crystal substrate for each equivalent defect orientation of the anisotropic defect.

[0072] For example, in the case of the rod-shaped defect 70 in the silicon single crystal substrate 80 having the (100) surface, where the major axis direction of the defect is equivalent to the

[0110] direction, the defect orientations with an aspect ratio of the surface shape of 1 to a are

[0011] and [0-11]. Also, the defect orientations with an aspect ratio of the surface shape of 1 to 2^0.5 are

[0110] ,

[0101] , [1-10], and [10-1].

[0073] Next, from the three-dimensional shape and aspect ratio of the anisotropic defect, the ratio of the depth is obtained for each equivalent defect orientation. For the rod-shaped defect 70 in the silicon single crystal substrate 80 with the (100) plane, where the major axis direction of the defect is equivalent to the

[0110] direction, when the defect orientation is

[0011] and [0-11], the major axis of the defect is parallel to the surface, and the depth is 1, which is the length in the minor axis direction. Also, when the defect orientation is

[0110] ,

[0101] , [1-10], and [10-1], the defect orientation forms an angle of 45° with the surface of the substrate 80. Therefore, the depth is a×sin45° = a×2^-0.5. That is, the ratio of the depth is 1 to a×2^-0.5.

[0074] Next, calculate the product of the ratio of the number of equivalent defect orientations and the ratio of the depth for each equivalent defect orientation, and determine the abundance of the surface shape on the substrate for each equivalent crystal orientation of the anisotropic defect. In the case of the rod-shaped defect 70, the ratio of the surface shape with an aspect ratio of 1:a to the surface shape with an aspect ratio of 1:√2 is 2×1:4×(a×2^-0.5)=1:a×√2.

[0075] This is to determine the abundance ratio according to the detection size, taking into account the phenomenon that, for defects with a longer depth in the defect orientation, there are more defects on the substrate surface than those with a shorter length in the depth direction of the defect orientation, even with the same number of defects.

[0076] In addition, when there are multiple types of anisotropic defects on the substrate to be evaluated, calculate the abundance ratio of the surface shape for each type of anisotropic defect.

[0077] (Step of estimating the number of anisotropic defects having the surface shape of the maximum detection size (Step 5)) Next, estimate the number of anisotropic defects having the surface shape of the maximum detection size present on the surface of the silicon single crystal substrate.

[0078] When all the foreign matters detected in the first step are observed in the second step, among the crystal defects observed in the second step, the number of anisotropic defects having the surface shape evaluated as having the maximum detection size in the third step is the number of anisotropic defects of the maximum detection size.

[0079] When a randomly selected part of the foreign matters detected in the first step is observed in the second step, first count the number of anisotropic defects having the surface shape evaluated as having the maximum detection size in the third step among the crystal defects observed in the second step.

[0080] Next, multiply the number of the anisotropic defects by the total number of the foreign matters detected in the first step and divide the result by the number of the foreign matters observed in the second step, and estimate it as the number of anisotropic defects of the maximum detection size present on the substrate surface.

[0081] Thus, it becomes possible to estimate the number of anisotropic defects of the maximum detectable size present on the entire substrate surface by assuming that the number of anisotropic defects of the maximum detectable size contained in foreign matter not observed in the second step exists at the same ratio as the foreign matter observed in the second step.

[0082] In addition, when there are multiple types of anisotropic defects on the substrate to be evaluated, the number of anisotropic defects of the maximum detectable size is estimated for each type of anisotropic defect.

[0083] (Step of estimating the number of anisotropic defects having a surface shape with a size smaller than the maximum detectable size (Step 6)) Next, from the abundance ratio calculated in the fourth step and the number of anisotropic defects having the surface shape of the maximum detectable size estimated in the fifth step present on the silicon single crystal substrate surface, the number of anisotropic defects having a surface shape with a size smaller than the maximum detectable size is estimated.

[0084] Specifically, it can be estimated that the number of anisotropic defects having the surface shape of the maximum detectable size multiplied by the abundance ratio of the anisotropic defects having a surface shape smaller than the maximum detectable size to the anisotropic defects having the surface shape of the maximum detectable size is the number on the substrate surface of the anisotropic defects having a surface shape smaller than the maximum detectable size.

[0085] Thus, it becomes possible to estimate the number of small defects with a surface shape that is difficult to detect by the surface inspection apparatus.

[0086] Also, for a substrate inspected by the surface inspection apparatus with a relatively large detection lower limit size, it becomes possible to estimate the number of defects with a relatively small detection size. To reduce the detection lower limit size of the surface inspection apparatus, a long inspection time is required. Therefore, by applying the evaluation method of the present invention, it becomes possible to estimate the number of defects with a small detection size in a shorter inspection time.

[0087] That is, when measured at the detection lower limit size of the surface inspection apparatus and using the evaluation method of the present invention, it becomes possible to estimate the number of defects smaller than the detection lower limit size of the surface inspection apparatus.

[0088] According to the method for evaluating crystal defects of such a silicon single crystal substrate of the present invention described above, the number of high aspect ratio crystal defects present on the substrate surface can be evaluated regardless of the type of defects, and defects having a size less than the detection limit size of the surface inspection apparatus can also be evaluated.

Example

[0089] Hereinafter, the present invention will be specifically described with reference to examples, which do not limit the present invention.

[0090] (Example) (Step of obtaining position information of foreign matter (first step)) First, as a substrate to be evaluated, a mirror-polished silicon single crystal substrate 80 having a diameter of 300 mm and a (110) plane as the main plane, which was cut out from a silicon single crystal doped with nitrogen by the Czochralski method, was prepared and washed. It was confirmed by a known defect inspection method by etching that the substrate 80 contains many COP (Crystal Originated Particle).

[0091] Next, the entire surface of the substrate 80 was measured with a surface inspection apparatus (SP7, Oblique mode (low angle incidence), DW1 channel, 19 nm up), and the coordinates of foreign matter were obtained. As a result, 5,177 foreign matters were detected.

[0092] (Step of classifying foreign matter into crystal defects and foreign matter other than crystal defects and evaluating the type of crystal defects (second step)) Next, based on the coordinates of the foreign matter obtained in the first step, the surface shape of all the foreign matter was observed with a review SEM (eDR7280), and the types of crystal defects among the observed foreign matter were classified. The types of crystal defects were three known types of crystal defects: rod-shaped COP with a major axis in the <110> direction, which is common in nitrogen-doped silicon crystals, flat-plate-shaped COP, which is also common in nitrogen-doped silicon crystals, and regular octahedron-shaped COP, which is also observed when nitrogen doping is not performed.

[0093] (Step (the third step) of verifying whether it is an anisotropic defect in which the detection size of the crystal defect varies depending on the defect orientation of the crystal defect and evaluating the surface shape of the maximum detection size) First, among the three types of crystal defects, the flat COP and the octahedral COP have a low aspect ratio, so it was evaluated that the detection size does not change significantly depending on the defect orientation.

[0094] Next, for each form of the defect orientation ([-110],

[0101] ,

[0011] , [01-1], [10-1],

[0110] ) of the rod-shaped COP with the major axis in the <110> direction, it was evaluated whether the detection sizes are different from the surface shape on the (110) plane.

[0095] Generally, COP has a three-dimensional shape surrounded by {111} planes, but for simplicity, the three-dimensional shape was approximated by a quadrangular prism.

[0096] For the rod-shaped COP90 with a defect orientation in the [-110] direction with respect to the (110) plane, as shown in FIG. 6, the defect orientation is parallel to the surface. Therefore, when the length of the minor axis (the direction perpendicular to the major axis) is 1 and the length of the major axis is a, as shown in FIG. 7, the surface shape (surface shape A) is long in the [-110] direction with a length of a in the [-110] direction and a length of 1 in the

[0001] direction. Also, from the observation result of the review SEM shown in FIG. 8, it was confirmed that a is 2 on the substrate 80.

[0097] For rod-shaped COP90 with defect orientations in the forms of

[0101] ,

[0011] , [01-1], and [10-1] with respect to the (110) plane, as shown in Fig. 9, they intersect the (110) plane at 30°. Also, when projecting the

[0101] ,

[0011] , [01-1], and [10-1] directions onto the (110) plane, they become [1-12], [-112], [-11-2], and [1-1-2] respectively. Here, considering the rod-shaped COP90 with a defect orientation of

[0101] , when the length of the minor axis is 1, the length in the [1-12] direction is at most approximately 1 / sin30° = 2. Therefore, these rod-shaped COP90 have a shape (surface shape B) that is long in the [1-12] direction with a length of 2 in the [1-12] direction and a minor axis length of 1, as shown in Fig. 10.

[0098] The rod-shaped COP90 with a defect orientation of

[0110] with respect to the (110) plane is perpendicular to the surface as shown in Fig. 11, so it has a surface shape (surface shape C) with a length of 1 in the [-110] direction and the

[0001] direction as shown in Fig. 12.

[0099] From the above results, the surface shape aspect ratio of the rod-shaped COP90 on the substrate 80 is 1:2 for the rod-shaped COP90 with defect orientations of [-110],

[0101] ,

[0011] , [01-1], and [10-1], and 1:1 for the rod-shaped COP90 with a defect orientation of

[0110] . It was determined that they are anisotropic defects with different detection sizes depending on the form of the defect orientation.

[0100] Next, for the rod-shaped COP90, the surface shape with the maximum detection size was evaluated. As described above, the rod-shaped COP90 with a defect orientation of [-110] and the rod-shaped COP90 with a defect orientation of

[0101] ,

[0011] , [01 - 1], and [10 - 1] rod - shaped COP90s have different defect orientations. When the length of the minor axis in the three - dimensional shape is set to 1, the surface geometric area is 1×2 and is equivalent. On the other hand, the rod - shaped COP90 with a defect orientation of

[0110] has an area of 1×1, which is smaller than other defect orientations. From the above results, for the rod - shaped COP90 on the substrate 80, the surface shapes with the largest detection size are evaluated to be two types: when the defect orientation is [-110] (the length in the [-110] direction is 2 and the length in the

[0001] direction is 1), and when the defect orientations are

[0101] ,

[0011] , [01 - 1], and [10 - 1] (the length in the [1 - 12] direction is 2 and the length in the [-111] direction is 1).

[0101] (Step (the fourth step) of calculating the abundance ratio of the surface shape on the silicon single - crystal substrate surface) Next, for the rod - shaped COP90, the number of forms of the defect orientation taking the surface shape for each surface shape was determined. When the surface shape is such that the length in the [-110] direction is 2 and the length in the

[0001] direction is 1 (surface shape A), the form of the defect orientation taking this surface shape is one form of [-110]. When the surface shape is such that the length in the [1 - 12] direction is 2 and the length in the [-111] direction is 1 (surface shape B), the forms of the defect orientation taking this surface shape are four forms:

[0101] ,

[0011] , [01 - 1], and [10 - 1]. When the surface shape is such that the lengths in the [-110] direction and the

[0001] direction are 1 (surface shape C), the form of the defect orientation taking this surface shape is one form of

[0110] .

[0102] Next, from the three - dimensional shape and aspect ratio of the rod - shaped COP90, the ratio of the depth for each form of the defect orientation was determined. In the case of the form with a defect orientation of [-110] (corresponding to surface shape A), since the defect orientation is parallel to the (110) surface, the depth coincides with the length of the minor axis, which is 1. In the case of the forms with defect orientations of

[0101] ,

[0011] , [01 - 1], and [10 - 1] (corresponding to surface shape B), the defect orientation is at an angle of 30° with respect to the surface. Therefore, the approximate depth is the length of the major axis, 2×sin30° = 1 (the central region of the defect where the detection size is detected larger than the actual size). In the case of the form with a defect orientation of [-110] (corresponding to surface shape C), the defect orientation is perpendicular to the surface, and the depth coincides with the length of the major axis, which is 2.

[0103] When the abundance ratios of the forms of these defect orientations are the same in the silicon single crystal, the forms of the defect orientations with a long length in the depth direction are likely to be exposed on the surface of the cut-out substrate. Since the probability of their existence increases in proportion to the length in the depth direction, the number of defect forms with surface shape C is about twice that on the substrate surface compared to the number of forms of each defect orientation with surface shapes A and B.

[0104] From the above results, the abundance ratios on the surfaces of the substrates 80 with surface shapes A, B, and C were 1×1:4×1:1×2 = 1:4:2.

[0105] (Step (the fifth step) of estimating the number of anisotropic defects having the surface shape with the maximum detection size) Next, among the rod-shaped COP90s observed in the second step, the number of those with a surface shape having a length of 2 in the [-110] direction and a length of 1 in the

[0001] direction (surface shape A), and those with a surface shape having a length of 2 in the [1-12] direction and a length of 1 in the [-111] direction (surface shape B), which were evaluated to have the maximum detection size in the third step, was investigated. As a result, the total of both was 1480.

[0106] (Step (the sixth step) of estimating the number of anisotropic defects having a surface shape with a size smaller than the maximum detection size) Next, from the results of each step so far, the number of crystal defects having a surface shape with a relatively small detection size among the surface shapes of the rod-shaped COP90s, with a length of 1 in both the [-110] direction and the

[0001] direction (surface shape C), present on the surface of the substrate 80 was estimated. In the fifth step, 1480 rod-shaped COP90s with the surface shape having the maximum detection size were present on the substrate surface. On the other hand, the surface abundance ratio between the surface shape having the maximum detection size and the surface shape with a length of 1 in both the [-110] direction and the

[0001] direction was (1 + 4):2 = 5:2. Therefore, it was estimated that about 1480×2 / 5 ≒ 714 rod-shaped COP90s with a surface shape having a length of 1 in both the [-110] direction and the

[0001] direction were present on the surface.

[0107] (Comparative example) Regarding the substrate 80 used in the example, among the crystal defects classified in the second step of the example, the number of rod-shaped COPs having a surface shape corresponding to the surface shape C in the example was counted. As a result, the number of rod-shaped COPs having a surface shape corresponding to the surface shape C in the example was 586. This is less than the 714 estimated in the example.

[0108] As described above, according to the example of the present invention, the number of defects in a surface shape with a relatively small detection size can be estimated more accurately than when measured at the minimum detection lower limit size of the surface inspection apparatus.

[0109] (Reference Example) The substrate 80 used in the example was measured by a surface inspection apparatus (SP7, Oblique mode (low angle incidence), DW1 channel) while changing the detection lower limit size in the range from 19 nm up to 25 nm up, and the number of defects in the surface shape C was counted in the same manner as in the comparative example for each. When the number of defects in the surface shape C is plotted against the detection lower limit size, it becomes as shown in FIG. 13. When the approximate straight line is extrapolated to the side where the detection lower limit size is small, the 714 estimated in the example corresponds to a detection lower limit size of 16 nm up. Since the highest sensitivity detection lower limit size applicable to the substrate 80 is 19 nm up, it was shown that by the method of the present invention, for defects in a surface shape with a relatively small detection size, the number of defects smaller than the detection lower limit size can be estimated.

[0110] This specification includes the following aspects. [1]: A first step of measuring a mirror-polished single-crystalline silicon substrate with a surface inspection apparatus, detecting foreign matter on the single-crystalline silicon substrate, and acquiring position information of the foreign matter; observing the surface shape of the foreign matter based on the position information of the foreign matter, classifying the foreign matter into crystal defects and foreign matter other than the crystal defects, and evaluating the type based on the shape of the crystal defects; a second step; for each type based on the shape of the crystal defects, verifying whether the detected size of the crystal defect is an anisotropic defect that varies depending on the defect orientation of the crystal defect, and for the crystal defects evaluated as the anisotropic defects, for each type of the anisotropic defects, evaluating the surface shape with the maximum detected size among the surface shapes on the surface of the single-crystalline silicon substrate for each equivalent defect orientation of the anisotropic defects; a third step; for each type of the anisotropic defects, calculating the abundance ratio of the surface shape on the surface of the single-crystalline silicon substrate for each equivalent defect orientation of the anisotropic defects from the three-dimensional shape and orientation of the anisotropic defects; a fourth step; estimating the number of the anisotropic defects having the surface shape with the maximum detected size present on the surface of the single-crystalline silicon substrate; a fifth step; and estimating the number of the anisotropic defects having a surface shape smaller than the maximum detected size from the abundance ratio calculated in the fourth step and the number of the anisotropic defects having the surface shape with the maximum detected size present on the surface of the single-crystalline silicon substrate estimated in the fifth step. A method for evaluating crystal defects of a single-crystalline silicon substrate including a sixth step. [2]: The method for evaluating crystal defects of the single-crystalline silicon substrate according to [1] above, wherein the second step includes observing the surface shape of the foreign matter using a microscope. [3]: The method for evaluating crystal defects of the single-crystalline silicon substrate according to [1] or [2] above, wherein the second step includes observing the shape of a cross-section of the crystal defect perpendicular to the surface of the single-crystalline silicon substrate that intersects the crystal defect using a transmission electron microscope or a scanning transmission electron microscope. [4]: The method for evaluating crystal defects of the single-crystalline silicon substrate according to [1], [2], or [3] above, wherein the third step includes evaluating based on the three-dimensional shape of a known crystal defect. [5]: The method for evaluating crystal defects of the single-crystalline silicon substrate according to [3] or [4] above, wherein the third step includes evaluating based on the three-dimensional shape of the crystal defect evaluated by observing the shape of the cross-section.

[0111] Note that the present invention is not limited to the above-described embodiments. The above-described embodiments are examples, and any configuration that has substantially the same configuration as the technical idea described in the claims of the present invention and exhibits the same operational effects is included in the technical scope of the present invention.

Explanation of Reference Numerals

[0112] 70... rod-shaped defect, 80... single-crystal silicon substrate, 90... rod-shaped COP.

Claims

1. A first step of measuring a mirror-polished single-crystal silicon substrate with a surface inspection device, detecting foreign matter on the single-crystal silicon substrate, and acquiring position information of the foreign matter; A second step of observing the surface shape of the foreign matter based on the position information of the foreign matter, classifying the foreign matter into crystal defects and foreign matter other than the crystal defects, and evaluating the type based on the shape of the crystal defects; For each type based on the shape of the crystal defects, verifying whether the detected size of the crystal defects is an anisotropic defect that varies depending on the defect orientation of the crystal defects, and for the crystal defects evaluated as the anisotropic defects, for each type of the anisotropic defects, evaluating the surface shape with the maximum detected size among the surface shapes on the surface of the single-crystal silicon substrate for each equivalent defect orientation of the anisotropic defects; A fourth step of calculating, for each type of the anisotropic defects, the abundance ratio of the surface shape on the surface of the single-crystal silicon substrate for each equivalent defect orientation of the anisotropic defects from the three-dimensional shape and orientation of the anisotropic defects; A fifth step of estimating the number of the anisotropic defects having the surface shape with the maximum detected size present on the surface of the single-crystal silicon substrate; A sixth step of estimating the number of the anisotropic defects having a surface shape smaller than the maximum detected size from the abundance ratio calculated in the fourth step and the number of the anisotropic defects having the surface shape with the maximum detected size present on the surface of the single-crystal silicon substrate estimated in the fifth step. A method for evaluating crystal defects of a single-crystal silicon substrate, characterized by including the above steps.

2. The second step includes observing the surface shape of the foreign matter using a microscope. The method for evaluating crystal defects of a single-crystal silicon substrate according to Claim 1.

3. The second step includes observing the shape of a cross-section of the crystal defect perpendicular to the surface of the single-crystal silicon substrate that intersects the crystal defect using a transmission electron microscope or a scanning transmission electron microscope. The method for evaluating crystal defects of a single-crystal silicon substrate according to Claim 2.

4. The evaluation method of crystal defects of a silicon single crystal substrate according to any one of claims 1 to 3, wherein the third step includes evaluating based on the three-dimensional shape of known crystal defects.

5. The evaluation method of crystal defects of a silicon single crystal substrate according to claim 3, wherein the third step includes evaluating based on the three-dimensional shape of the crystal defects evaluated by observing the shape of the cross section.

Images