Defect evaluation method for semiconductor silicon wafer

Abstract

The present invention is a defect evaluation method for a semiconductor Si wafer for evaluating a defect shape inside the Si wafer, the defect evaluation method comprising: a first defect detection step for irradiating a surface of an Si wafer with a DUV laser beam to acquire the position coordinates of defects included in the outermost surface; a second defect detection step for irradiating the surface of the Si wafer with a visible-light laser beam to acquire the position coordinates of defects included in a surface layer region including the outermost surface; a defect classification step for comparing the position coordinates of the defects acquired in the first and second defect detection steps, classifying defects detected only in the first defect detection step and defects of the same coordinates among the defects detected in the first and second defect detection steps into exposed defects, and classifying defects detected only in the second defect detection step into non-exposed defects; and a defect observation step for observing the shapes of the defects classified into the non-exposed defects. Accordingly, a method for observing the overall situations of defects present in an Si wafer without destroying the structures of the defects is provided.

Description

Semiconductor silicon wafer defect evaluation method

[0001] The present invention relates to a method for evaluating defects in a semiconductor silicon wafer.

[0002] Semiconductor devices are used for a variety of purposes, and the type and quality of silicon substrates required vary depending on the device's performance and application. Silicon manufacturers are therefore required to be able to respond flexibly to these demands, making defect control technology important in silicon single crystal production and wafer processing. To improve defect control technology, it is essential to accurately detect defects formed during silicon single crystal production and wafer processing, understand their true nature, and clarify their formation mechanisms.

[0003] Defects that exist near the wafer surface are generally detected using a surface defect inspection device (particle counter), while defects that exist inside the wafer are generally detected using infrared tomography.

[0004] These devices irradiate the wafer surface with laser light and detect the light scattered from defects with a detector. Surface and internal defects are detected by utilizing the fact that the penetration depth of light changes depending on the wavelength of the irradiating laser.

[0005] For example, Patent Document 1 discloses a technique in which light of two wavelengths whose absorption coefficients for silicon differ by about one order of magnitude is irradiated onto a wafer surface, the intensity of scattered light from defects is measured for each wavelength, and the depth position and size of the defects are determined. Patent Document 2 discloses a technique in which light of two wavelengths whose penetration depths differ by three times or more is scanned across a sample and irradiated onto a wafer, the scattered light from internal defects is measured for each wavelength, the defect size is derived from the scattered light intensity on the longer wavelength side, the depth position is derived from the ratio of the scattered light intensities, and the depth position and size are displayed as the wafer in-plane distribution of the defects.

[0006] Japanese Patent Application Laid-Open No. 11-237226 International Publication No. 97 / 35162

[0007] To improve defect control technology, it is necessary to understand the true nature of defects. To do this, it is essential to understand the overall picture of defects. For example, infrared tomography is known to involve cleaving a wafer, irradiating light from the wafer surface, and detecting scattered light from the cleaved surface. However, while this method can determine the defect position (X, Y) as seen from the cleaved surface, it is difficult to determine the distance (Z) from the cleaved surface to the defect, and it is also difficult to observe defects using other methods.

[0008] On the other hand, particle counters with a laser wavelength of 266 nm (e.g., KLA-Tencor's Surfscan SP7) can detect only defects near the outermost surface because the laser barely penetrates the silicon, and can also acquire defect coordinates, making defect observation easy. However, among the defects detected, crystal defects (e.g., COPs), excluding processing-induced defects such as foreign matter and scratches adhering during the wafer processing process, are detected by particle counters because a portion of the defect is cut off during the wafer processing process and exposed on the surface. In other words, even if these surface-exposed defects are observed using a microscope or the like, only a portion cut off from the defect formed during crystal manufacturing, i.e., a defective structure with a defect, is observed. Therefore, in order to observe the entire picture of the defects, it is important to detect and extract defects not exposed on the wafer surface (non-exposed defects), and this observation must be performed without destroying the defect structure.

[0009] Patent Documents 1 and 2 describe techniques for identifying the depth and position of a defect from optical information obtained by irradiating a laser beam, but do not describe a means for observing the entire image of the defect without destroying the structure.

[0010] The present invention has been made to solve the above problems, and an object of the present invention is to provide a method for extracting unexposed defects present in a semiconductor silicon wafer and observing the overall image of the defects without destroying the structure of the defects.

[0011] The present invention has been made to achieve the above-mentioned object, and provides a semiconductor silicon wafer defect evaluation method for evaluating the shapes of defects inside the semiconductor silicon wafer, the semiconductor silicon wafer defect evaluation method comprising: a first defect detection step of irradiating a surface of the semiconductor silicon wafer with a DUV laser and detecting scattered light from the outermost surface of the semiconductor silicon wafer to acquire position coordinates of defects contained in the outermost surface; a second defect detection step of irradiating the surface of the semiconductor silicon wafer with a visible light laser and detecting reflected light from a surface layer region including the outermost surface of the semiconductor silicon wafer to acquire position coordinates of defects contained in the surface layer region; a defect classification step of comparing the position coordinates of the defects acquired in the first and second defect detection steps, and classifying defects detected only in the first defect detection step and those having the same coordinates as exposed defects, and classifying defects detected only in the second defect detection step as non-exposed defects; and a defect observation step of observing the shapes of the defects classified as non-exposed defects.

[0012] According to this semiconductor silicon wafer defect evaluation method, only defects that are free from defects in the defect structure present in the semiconductor silicon wafer can be extracted as non-exposed defects, and the entire image of the defect can be observed without destroying the defect structure.

[0013] In this case, the wavelength of the DUV laser can be set to 266 nm.

[0014] At this wavelength, almost no light penetrates into the silicon wafer, making it possible to detect scattered light from defects on the top surface of the wafer with high accuracy.

[0015] In this case, the wavelength of the visible light laser can be set to 400 to 700 nm.

[0016] This allows for more stable and effective detection of reflected light from defects contained in the surface layer region including the outermost surface.

[0017] In this case, the observation in the defect observation step can be a cross-sectional observation of the defect classified as the non-exposed defect.

[0018] This allows for more accurate observation of the entire defect without destroying the structure of the defect that is not exposed on the surface.

[0019] In this case, the cross section can be observed by processing the periphery of the defect classified as the non-exposed defect with a focused ion beam, and observing the cross section with a transmission electron microscope or a scanning transmission electron microscope.

[0020] This allows the entire defect to be observed with higher precision without destroying the structure of the defect that is not exposed on the surface.

[0021] As described above, according to the semiconductor silicon wafer defect evaluation method of the present invention, only defects that are free from defects in the defect structure present in the semiconductor silicon wafer can be extracted as non-exposed defects, and the entire image of the defect can be observed without destroying the defect structure.

[0022] FIG. 1 shows a flowchart of an example of a semiconductor silicon wafer defect evaluation method according to the present invention; FIG. 2 shows a schematic diagram of a measurement device that can be used in a first defect detection step; FIG. 3 shows a schematic diagram of a measurement device that can be used in a second defect detection step; FIG. 4 shows the results of cross-sectional TEM observation of an example; FIG. 5 shows other results of cross-sectional TEM observation of an example; and FIG. 6 shows the results of cross-sectional TEM observation of comparative example 1.

[0023] The present invention will be described in detail below, but the present invention is not limited thereto.

[0024] As described above, there has been a demand for a method for extracting unexposed defects present in a semiconductor silicon wafer (hereinafter simply referred to as a "silicon wafer" or "wafer") and observing the overall image of the defects without destroying the structure of the defects.

[0025] As a result of extensive research into the above-mentioned problems, the present inventors have discovered a defect evaluation method for semiconductor silicon wafers for evaluating defect shapes inside the semiconductor silicon wafers, the defect evaluation method for semiconductor silicon wafers comprising: a first defect detection step of irradiating a surface of the semiconductor silicon wafer with a DUV laser and detecting scattered light from the outermost surface of the semiconductor silicon wafer to obtain position coordinates of defects contained in the outermost surface; a second defect detection step of irradiating the surface of the semiconductor silicon wafer with a visible light laser and detecting reflected light from a surface layer region including the outermost surface of the semiconductor silicon wafer to obtain position coordinates of defects contained in the surface layer region; The inventors have found that a semiconductor silicon wafer defect evaluation method comprising a defect classification step of comparing the position coordinates of the defects acquired in the first and second defect detection steps, and classifying defects detected only in the first defect detection step and defects detected in the first and second defect detection steps that have the same coordinates as exposed defects, and classifying defects detected only in the second defect detection step as non-exposed defects, and a defect observation step of observing the shapes of the defects classified as non-exposed defects, can extract only defects that have no defects in the defect structure present in the semiconductor silicon wafer as non-exposed defects, and can observe the overall image of the defects without destroying the defect structure, and have completed the present invention.

[0026] A flowchart of an example of a semiconductor silicon wafer defect evaluation method according to the present invention is shown in Figure 1. First, as shown in S1 of Figure 1, a silicon wafer that has been mirror-polished by a general method is prepared.

[0027] [First defect detection process] Next, as shown in S2 of FIG. 1 , a particle counter equipped with a DUV laser irradiates the surface of the silicon wafer with a DUV laser to scan it, and detects scattered light containing optical information from defects on the top surface of the silicon wafer to obtain position coordinates of defects contained in the top surface (first defect detection process).

[0028] DUV is ultraviolet light having a wavelength of about 200 to 300 nm. In particular, the wavelength of a DUV laser can be 266 nm.

[0029] At this wavelength, almost no light penetrates into the silicon wafer, making it possible to detect scattered light from defects on the wafer's top surface with high accuracy.

[0030] In the first defect detection step, in order to detect defects present on the outermost surface, it is preferable to use an oblique incidence device which has a shallower penetration depth of light. More preferably, the incident angle can be set to 30° or less with respect to the wafer surface.

[0031] 2 shows a schematic diagram of a measurement device that can be used in the first defect detection step. A detector 1 is installed above a semiconductor silicon wafer 4. The detector 1 detects scattered light 3 from the outermost surface of the semiconductor silicon wafer 4, which is irradiated with DUV laser light 2, and obtains the position coordinates of defects contained in the outermost surface.

[0032] [Second Defect Detection Step] Next, as shown in S3 of Fig. 1, a particle counter equipped with a visible laser irradiates and scans the surface of the silicon wafer with a visible laser, and detects reflected light containing optical information from defects in the surface region, including the outermost surface of the silicon wafer, to obtain position coordinates of defects contained in the surface region (second defect detection step). Here, the surface region includes the outermost surface and the region from the outermost surface in the depth direction. The depth depends on the wavelength of the visible laser.

[0033] Visible light has a wavelength of 360 to 400 nm (lower limit) and 760 to 830 nm (upper limit). The wavelength of the visible laser may be within this range, preferably 400 to 700 nm. In this case, the depth of the surface layer is approximately 0.1 to 5 μm.

[0034] This allows for more stable and effective detection of reflected light from defects in the surface layer, including the outermost surface. The shorter the wavelength, the smaller the minimum detectable particle size but the shorter the penetration depth, while the longer the wavelength, the larger the minimum detectable particle size but the longer the penetration depth, allowing for detection of deeper defects.

[0035] The wavelength of the visible laser can be determined according to the size and depth of the defect to be detected, but is preferably set to 532 nm, which allows defects present at a depth of about 1 μm from the outermost surface to the surface layer to be detected.

[0036] On the other hand, in the second defect detection process, since the purpose is to detect defects present in the surface region including the outermost surface, the incident angle is not particularly limited, but it is preferable to use an apparatus with normal incidence or a small incident angle.

[0037] 3 shows a schematic diagram of a measurement device that can be used in the second defect detection step. A detector 5 is installed above the semiconductor silicon wafer 4. Visible laser light 6 is irradiated onto the surface of the semiconductor silicon wafer 4, and reflected light 7 from a surface region including the outermost surface is detected by the detector 5 to obtain the position coordinates of defects contained in the surface region.

[0038] Therefore, it is preferable to use different devices for the first and second defect detection steps. In such a preferred embodiment, since the measurement devices used in the first and second defect detection steps are different as shown in Figures 2 and 3, the defect coordinates of the same defect may not necessarily match. In such a case, all defects within a specific coordinate range can be considered to be the same defect. The specific coordinate range is not particularly limited, but is preferably 50 to 200 μm, taking into account the coordinate accuracy of the substrate surface inspection device. Furthermore, coordinate correction between the devices may be performed using a sample with known defect coordinates.

[0039] This allows foreign matter attached to the surface, defects exposed on the surface that are partially missing, and processing-induced defects formed on the surface to be excluded as exposed defects from the defects detected in the first and second defect detection processes, and therefore allows defects that are not exposed on the surface and do not have missing defect structures to be extracted with high accuracy as non-exposed defects.

[0040] 1 , the position coordinates of the defects acquired in the first and second defect detection processes are compared, and defects detected only in the first defect detection process and defects detected in the first and second defect detection processes that have the same coordinates are classified as exposed defects, and defects detected only in the second defect detection process are classified as non-exposed defects (defect classification process). In other words, exposed defects are defects detected in the first defect detection process, and non-exposed defects are defects detected in the second defect detection process, excluding defects detected in the first and second defect detection processes that have the same coordinates.

[0041] 1, the shapes of the defects classified as non-exposed defects are observed (defect observation step). This makes it possible to extract and observe only defects that have no defects in their defect structure.

[0042] In this case, the observation in the defect observation step can be a cross-sectional observation of the defects classified as non-exposed defects, which allows the overall image of the defect to be observed with higher precision without destroying the structure of the defect that is not exposed on the surface.

[0043] In addition, the cross-section observation can be performed by processing the area around a defect classified as a non-exposed defect with a focused ion beam (FIB) to form a cross-section, and observing the cross-section with a transmission electron microscope (TEM) or a scanning transmission electron microscope (STEM).

[0044] It is preferable to first prepare a relatively large TEM sample, confirm the presence of defects through TEM observation, and then gradually proceed with thinning, which allows for non-destructive and highly accurate observation of the entire image of unexposed defects.

[0045] Furthermore, when defect observation is performed using these devices, elemental analysis and structural analysis may be performed using elemental analysis techniques associated with the devices, such as energy dispersive X-ray spectroscopy (EDX), electron energy loss spectroscopy (EELS), and electron beam diffraction, which allows for more detailed defect analysis.

[0046] The present invention will be specifically explained below with reference to examples, but the present invention is not limited to these examples.

[0047] Example First, a silicon wafer having a diameter of 300 mm was prepared, which had been mirror-polished by a common method.

[0048] Next, the wafer was measured (first defect detection step) in oblique (grazing incidence) mode / 15 nm up (detection size≧15 nm) using a Surfscan SP7 (wavelength 266 nm) manufactured by KLA-Tencor Corporation.

[0049] Next, the wafer was measured (second defect detection step) using a MAGICS M5640 (wavelength 532 nm) manufactured by Lasertec Corporation at a slice level of 40 mV (corresponding to a detection size of 50 to 60 nm or more).

[0050] Next, the defect coordinates obtained by the two measurements were compared, and defects detected only in the first defect detection process and defects with the same coordinates among defects detected in the first and second defect detection processes were classified as exposed defects, and defects detected only in the second defect detection process were classified as non-exposed defects.

[0051] Next, two defects were randomly selected from the non-exposed defects, processed with a focused ion beam (FIB), and cross-sectional TEM observation was performed. As a result, non-exposed defects 8 were observed, as shown in Figures 4 and 5, and the entire image of the defects not exposed on the surface could be observed for all of the extracted non-exposed defects. In this way, it was confirmed that the entire image of the defect can be observed by performing cross-sectional observation of the non-exposed defects.

[0052] Comparative Example 1: A defect classified as an exposed defect in the example was processed with an FIB and subjected to cross-sectional TEM observation. As a result, an exposed defect 9 was observed as shown in Figure 6. Because a portion of the defect was cut away and exposed on the surface, it was not possible to obtain an overall image of the defect.

[0053] Comparative Example 2 The same silicon wafer as in the example was subjected to measurement (second defect detection step) using a Lasertec MAGICS M5640 (wavelength 532 nm) at a slice level of 40 mV (corresponding to a detection size of 50 to 60 nm or more).

[0054] Next, several defects detected in the above measurement were extracted, processed using FIB, and cross-sectional TEM observation was performed. However, it was found that defects exposed on the surface and defects not exposed on the surface were mixed together, making it difficult to accurately grasp the overall picture of the defects.

[0055] As described above, according to the embodiment of the present invention, only defects that do not have any defects in the defect structure present in the silicon wafer can be extracted as non-exposed defects, and the entire image of the defect can be observed without destroying the defect structure.

[0056] This specification includes the following aspects: [1]: A semiconductor silicon wafer defect evaluation method for evaluating the shapes of defects inside the semiconductor silicon wafer, the semiconductor silicon wafer defect evaluation method comprising: a first defect detection step of irradiating a surface of the semiconductor silicon wafer with a DUV laser and detecting scattered light from the outermost surface of the semiconductor silicon wafer to acquire position coordinates of defects contained in the outermost surface, a second defect detection step of irradiating the surface of the semiconductor silicon wafer with a visible light laser and detecting reflected light from a surface layer region including the outermost surface of the semiconductor silicon wafer to acquire position coordinates of defects contained in the surface layer region, a defect classification step of comparing the position coordinates of the defects acquired in the first and second defect detection steps to classify defects detected only in the first defect detection step and defects having the same coordinates as exposed defects and classifying defects detected only in the second defect detection step as non-exposed defects, and a defect observation step of observing the shapes of the defects classified as non-exposed defects. [2]: The method for evaluating defects in a semiconductor silicon wafer according to [1] above, which includes setting the wavelength of the DUV laser to 266 nm. [3]: The method for evaluating defects in a semiconductor silicon wafer according to [1] above or [2] above, which includes setting the wavelength of the visible light laser to 400 to 700 nm. [4]: ​​The method for evaluating defects in a semiconductor silicon wafer according to [1] above, [2] above or [3] above, which includes performing the observation in the defect observation step as cross-sectional observation of a defect classified as a non-exposed defect. [5]: The method for evaluating defects in a semiconductor silicon wafer according to [4] above, which includes performing the cross-sectional observation by processing the periphery of a defect classified as a non-exposed defect with a focused ion beam, and observing the cross-section formed by this processing with a transmission electron microscope or a scanning transmission electron microscope.

[0057] The present invention is not limited to the above-described embodiments, which are merely examples, and anything that has substantially the same configuration as 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.

Claims

1. A semiconductor silicon wafer defect evaluation method for evaluating the shapes of defects inside a semiconductor silicon wafer, comprising: a first defect detection step of irradiating a surface of the semiconductor silicon wafer with a DUV laser and detecting scattered light from the outermost surface of the semiconductor silicon wafer to obtain position coordinates of defects contained in the outermost surface; a second defect detection step of irradiating the surface of the semiconductor silicon wafer with a visible laser and detecting reflected light from a surface region including the outermost surface of the semiconductor silicon wafer to obtain position coordinates of defects contained in the surface region; a defect classification step of comparing the position coordinates of the defects obtained in the first and second defect detection steps to classify defects detected only in the first defect detection step and defects with identical coordinates detected in the first and second defect detection steps as exposed defects, and classifying defects detected only in the second defect detection step as non-exposed defects; and a defect observation step of observing the shapes of the defects classified as non-exposed defects.

2. The semiconductor silicon wafer defect evaluation method according to claim 1, wherein the wavelength of the DUV laser is 266 nm.

3. The semiconductor silicon wafer defect evaluation method according to claim 1, wherein the wavelength of the visible light laser is 400 to 700 nm.

4. A semiconductor silicon wafer defect evaluation method according to any one of claims 1 to 3, characterized in that the observation in the defect observation step is a cross-sectional observation of the defects classified as the non-exposed defects.

5. A method for evaluating defects in semiconductor silicon wafers as described in claim 4, characterized in that the cross-section observation is performed by processing the periphery of the defect classified as the non-exposed defect with a focused ion beam to form a cross-section, and observing the cross-section with a transmission electron microscope or a scanning transmission electron microscope.

Images