Crystal defect observation device

Abstract

To provide a crystal defect observation device capable of easily classifying the kind of a crystal defect of a semiconductor wafer.SOLUTION: A crystal defect observation device which observes a crystal defect of a semiconductor wafer comprises: a light source 1 which is collimated into parallel light flux; a ring diaphragm 4 which converges emitted light from the light source into a ring shape; a condenser lens 5 which focuses the emitted light passing the ring diaphragm and irradiates one face of a semiconductor wafer W; an objective lens unit 8 which focuses transmitted light which is transmitted through the semiconductor wafer and emitted from the other face and makes it into parallel light flux again; a phase shift plate 9 which is disposed at an optically conjugate position with the ring diaphragm; an imaging lens 10 which images the transmitted light passing the phase shift plate; a polarizer 12 which is disposed between the light source and the condenser lens; and an analyzer 13 which is disposed between the phase shift plate and the imaging lens. The phase shift plate generates a phase difference between a ring-like portion where direct light in the transmitted light passes and the other portion where diffraction light passes.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a crystal defect observation device. [Background technology]

[0002] In recent years, semiconductor devices have been used in a variety of fields. These semiconductor devices are formed by laminating insulating films, electrodes, and the like on a semiconductor wafer.

[0003] The performance of a semiconductor device is primarily determined by the characteristics of the semiconductor wafer itself. If a semiconductor wafer contains crystal defects, these defects can, for example, hinder the movement of electrons and holes within the semiconductor, preventing the semiconductor wafer from exhibiting its inherent characteristics. As a result, semiconductor devices constructed in areas of the semiconductor wafer that contain crystal defects may not exhibit the desired characteristics. Therefore, when manufacturing semiconductor devices, the semiconductor wafer that serves as the substrate is first inspected for crystal defects. If crystal defects are found, they are then observed to determine the type and state of the crystal defects.

[0004] A known device for observing these crystal defects is a crystal defect observation device that uses phase contrast observation of light passing through a semiconductor wafer (see JP 2021-19048 A). This crystal defect observation device directly observes changes in the refractive index of crystal defects by using phase contrast observation of light passing through a semiconductor wafer. This allows for accurate observation of crystal defects even in transparent semiconductor wafers. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Patent Publication No. 2021-19048 Summary of the Invention [Problem to be solved by the invention]

[0006] The conventional crystal defect observation device described above can effectively detect screw dislocations and edge dislocations among crystal defects. However, because the conventional crystal defect observation device simultaneously observes screw dislocations and edge dislocations, it is necessary to distinguish between dislocations based on their morphology, such as their appearance in an image. A known technique for distinguishing between dislocation types is to automatically identify them based on their characteristics through binarization analysis of dislocation contrast. However, dislocation contrast varies depending on the crystal quality of various manufacturing methods and their controllability, and the morphology of dislocation contrast varies depending on the quality, making identification difficult. While other methods, such as artificial intelligence (AI), require training data for learning, which in turn requires visual classification results. As such, the classification of crystal defects using the conventional crystal defect observation device described above tends to be time-consuming.

[0007] The present disclosure has been made in view of the above circumstances, and aims to provide a crystal defect observation device that can easily classify the types of crystal defects in semiconductor wafers. [Means for solving the problem]

[0008] A crystal defect observation device according to one embodiment of the present disclosure is a crystal defect observation device for observing crystal defects in a semiconductor wafer, and includes a light source collimated to a parallel beam of light, a ring aperture that narrows the light emitted from the light source into a ring shape, a condenser lens that focuses the light emitted after passing through the ring aperture and irradiates it onto one side of the semiconductor wafer, an objective lens that focuses the transmitted light that passes through the semiconductor wafer and emerges from the other side to form a parallel beam of light again, a phase shift plate arranged at an optically conjugate position with the ring aperture, an imaging lens that forms an image of the transmitted light that has passed through the phase shift plate, a polarizer arranged between the light source and the condenser lens, and an analyzer arranged between the phase shift plate and the imaging lens, wherein the phase shift plate generates a phase difference between a ring-shaped portion of the transmitted light through which direct light passes and another portion through which diffracted light passes. [Effects of the Invention]

[0009] The crystal defect observation device of the present disclosure can easily classify the types of crystal defects in semiconductor wafers. [Brief explanation of the drawings]

[0010] [Figure 1] FIG. 1 is a schematic diagram showing the configuration of a crystal defect observation device according to an embodiment of the present disclosure. [Figure 2] FIG. 2 is a schematic plan view showing the ring diaphragm of FIG. [Figure 3] FIG. 3 is a schematic plan view showing the phase shift plate of FIG. [Figure 4] FIG. 4 is a schematic diagram illustrating the principle of phase contrast microscopy. [Figure 5] FIG. 5 is a schematic perspective view showing an optical system for parallel polarized light of the crystal defect measuring device of FIG. [Figure 6] FIG. 6 is a schematic perspective view showing an optical system in orthogonal polarization of the crystal defect measuring device of FIG. DETAILED DESCRIPTION OF THE INVENTION

[0011] [Description of the embodiment of the present invention] The present inventors discovered that when observing light passing through a semiconductor wafer using phase contrast, superimposing polarized light changes the phase contrast of dislocations penetrating the crystal. They also discovered that specific crystal defects can be extracted by applying a specific polarization relationship, leading to the completion of the present disclosure.

[0012] That is, a crystal defect observation device according to one embodiment of the present disclosure is a crystal defect observation device for observing crystal defects in a semiconductor wafer, and comprises a light source collimated to a parallel beam of light, a ring aperture that narrows the light emitted from the light source into a ring shape, a condenser lens that focuses the light emitted after passing through the ring aperture and irradiates it onto one side of the semiconductor wafer, an objective lens that focuses the transmitted light that passes through the semiconductor wafer and emerges from the other side to form a parallel beam of light again, a phase shift plate arranged at an optically conjugate position with the ring aperture, an imaging lens that forms an image of the transmitted light that has passed through the phase shift plate, a polarizer arranged between the light source and the condenser lens, and an analyzer arranged between the phase shift plate and the imaging lens, and the phase shift plate generates a phase difference between a ring-shaped portion of the transmitted light through which direct light passes and another portion through which diffracted light passes.

[0013] This crystal defect observation device directly observes changes in the refractive index of crystal defects by observing light passing through a semiconductor wafer using phase contrast. Therefore, by using this crystal defect observation device, crystal defects can be accurately observed even in transparent semiconductor wafers. Furthermore, this crystal defect observation device can extract only specific crystal defects by establishing a specific polarization relationship between the polarizer and analyzer, making it easy to classify the types of crystal defects in semiconductor wafers.

[0014] Preferably, the polarizer is disposed between the light source and the annular aperture, which enables accurate detection of the effect of polarization modulation of transmitted light due to crystal defects in phase contrast observation, thereby enabling more sophisticated interpretation of information from phase contrast observation.

[0015] The polarization relationship between the polarizer and the analyzer can be switched between parallel polarization and orthogonal polarization. According to the present inventors, parallel polarization in a semiconductor wafer mainly extracts screw dislocations and edge dislocations, while orthogonal polarization mainly extracts screw dislocations. Therefore, by making the polarization relationship switchable between parallel polarization and orthogonal polarization, screw dislocations and edge dislocations can be easily classified among crystal defects in a semiconductor wafer.

[0016] [Details of the embodiment of the present invention] Hereinafter, a crystal defect observation device according to an embodiment of the present disclosure will be described in detail with reference to the drawings as appropriate.

[0017] The crystal defect observation device shown in FIG. 1 is a device for observing crystal defects in a semiconductor wafer W. The crystal defects that can be observed with this crystal defect observation device are all crystal defects in which the refractive index at the crystal defect location is different from that of the surrounding area. Examples of such crystal defects include threading screw dislocations (TSDs), threading edge dislocations (TEDs), basal plane dislocations (BPDs), and so-called composite dislocations, which are a combination of these dislocations. In particular, this crystal defect observation device can be suitably used for TSDs and TEDs, which are crystal dislocations that penetrate the main surface of the semiconductor wafer W in a direction perpendicular to the main surface.

[0018] As shown in Figure 1, the crystal defect observation device includes a light source 1, a reflecting mirror 2, a field diaphragm 3, a ring diaphragm 4, a condenser lens 5, an XY stage 6, an autofocus unit 7, an objective lens unit 8, a phase shift plate 9, an imaging lens 10, a camera 11, a polarizer 12, and an analyzer 13.

[0019] <Semiconductor wafer> Although the semiconductor wafer W to be measured is not particularly limited, the crystal defect observation device can be suitably used for inspecting hexagonal semiconductor wafers, particularly their C-faces. The semiconductor wafer W has birefringence and is transparent to visible light. Examples of such semiconductor wafers W include hexagonal semiconductor wafers, such as SiC (4H-SiC, 6H-SiC, 8H-SiC), GaN, and AlN.

[0020] In particular, semiconductor wafers under development, such as SiC, GaN, and AlN, have crystal defects with various crystal orientation anisotropies, unlike semiconductor wafers with extremely high single crystallinity, such as Si single crystal substrates. The crystal defect observation device can comprehensively observe crystal defects with various crystal orientation anisotropies and classify the types of crystal defects, making it particularly effective for such semiconductor wafers under development.

[0021] The diameter of the semiconductor wafer W is not necessarily limited, but the lower limit of the diameter of the semiconductor wafer W is preferably 0.5 inches, and more preferably 1 inch. On the other hand, the upper limit of the diameter of the semiconductor wafer W is preferably 8 inches, and more preferably 6 inches. If the diameter of the semiconductor wafer W is less than the lower limit, the observation cost for one semiconductor wafer W may become too high. Conversely, if the diameter of the semiconductor wafer W exceeds the upper limit, the crystal defect observation device will become larger, and the manufacturing cost of the crystal defect observation device may become too high.

[0022] The average film thickness of the semiconductor wafer W is not necessarily limited, but the lower limit of the average film thickness of the semiconductor wafer W is preferably 200 μm, more preferably 350 μm. On the other hand, the upper limit of the average film thickness of the semiconductor wafer W is preferably 1000 μm, more preferably 700 μm. If the average film thickness of the semiconductor wafer W is below the above lower limit, the semiconductor wafer W may be difficult to handle. On the other hand, if the average film thickness of the semiconductor wafer W exceeds the above upper limit, the transparency of the semiconductor wafer W may decrease, making it difficult to adjust the phase difference. Note that if the semiconductor wafer W is highly transparent, it can be observed with high accuracy even if the film thickness is greater than the above upper limit. Here, the "average film thickness" refers to the average value of film thicknesses measured at any 10 points.

[0023] The semiconductor wafer W typically has an epitaxial layer formed by epitaxially growing the same type of semiconductor on the surface of a bulk crystal. The crystal defect observation device can observe semiconductor wafers W having such an epitaxial layer on their surface, as well as semiconductor wafers W in their bulk crystal state (without an epitaxial layer).

[0024] <Light source> The light source 1 emits light to be irradiated onto the semiconductor wafer W. The light emitted from the light source 1 is collimated into a parallel beam.

[0025] The light source 1 has a relatively low directivity, for example, a directivity angle of 2θ 1 / 2 It is preferable to use a light source with a directional characteristic of 30° or more. For example, if a highly directional light source such as a laser light is used, the coherence becomes too high, and differences in brightness (interference fringes) are likely to occur due to factors other than crystal defects. This may make it difficult to observe crystal defects accurately.

[0026] As the light source 1, a known light source such as an LED or a halogen lamp can be used. The light emitted from the light source 1 is preferably light of a single wavelength. By making the emitted light light of a single wavelength, it becomes easier to adjust the phase difference. The light of a single wavelength may be obtained by selecting a light source 1 that emits light of a single wavelength, or by using a filter to obtain light of a single wavelength.

[0027] When the emitted light is light of a single wavelength, for example, light of 500 nm (green) may be used. However, if light of a wavelength of 500 nm is easily absorbed by the semiconductor wafer W, the intensity of the light will be attenuated as it passes through the semiconductor wafer W, and therefore it is necessary to use emitted light of another wavelength that will not be attenuated.

[0028] <Reflector> The reflecting mirror 2 adjusts the direction of the emitted light from the light source 1 so that the direction of travel of the light is parallel to the central axis M that passes through the center of the ring aperture 4, the condenser lens 5, the objective lens selected by the objective lens unit 8, and the phase shift plate 9.

[0029] Since it is preferable that the emitted light is parallel to the central axis M regardless of the position, the reflecting mirror 2 should be a concave mirror.

[0030] The reflecting mirror 2 can also be omitted. In this case, the orientation of the light source 1 is adjusted so that the light emitted from the light source 1 is parallel to the central axis M. However, it is preferable to use the reflecting mirror 2 because it is easier to adjust.

[0031] <Field of view diaphragm> Field diaphragm 3 blocks light that does not contribute to observation out of the outgoing light reflected by reflecting mirror 2. Field diaphragm 3 is a plate-like body that can block light, and can be configured to have a through-hole in the center that allows the outgoing light to pass through.

[0032] The size of the field stop 3 (the size of the plate-like body) is set to a size that can block the outgoing light reflected by the reflecting mirror 2 from passing through any place other than the through-hole. In other words, the size of the field stop 3 is set to be at least larger than the mirror surface of the reflecting mirror 2.

[0033] The through hole is preferably circular, and is arranged so that its center coincides with the central axis M. The diameter of the through hole is set to a size that allows irradiation of a ring hole 4a of the ring aperture 4, which will be described later, i.e., slightly larger than the outer diameter of the ring hole 4a.

[0034] The field stop 3 is not an essential component and can be omitted, but it is preferable to provide the field stop 3 from the viewpoint of preventing unwanted light rays from wandering around.

[0035] <Aperture ring> The diaphragm ring 4 narrows down the light emitted from the light source 1 into a ring shape. As shown in Fig. 2, the diaphragm ring 4 is a plate-like body that can block the light emitted from the light source 1 that has passed through the through-hole of the field stop 3, and has a ring-shaped hole 4a in the center that allows the emitted light to pass through.

[0036] The shape of the diaphragm ring 4 is not particularly limited, but can be, for example, circular as shown in Fig. 2. The size of the diaphragm ring 4 is determined so that light emitted from the field stop 3 does not leak out from its periphery, and when the diaphragm ring 4 is circular, it can have a diameter of, for example, 25 mm or more and 35 mm or less.

[0037] The ring hole 4a is configured in an annular shape. The ring diaphragm 4 also has a support beam 4b that crosses part of the ring hole 4a and supports the plate-like portion inside the ring hole 4a. It is preferable that this support beam 4b be configured to have the smallest area in a plan view as long as it is able to support the inner portion. In FIG. 2, thin support beams 4b are provided in three locations to support the inner portion, but the configuration of the support beams 4b is not limited to this.

[0038] The lower limit of the outer diameter of the ring hole 4a is preferably 1.5 mm, more preferably 1.7 mm. On the other hand, the upper limit of the outer diameter of the ring hole 4a is preferably 4.5 mm, more preferably 4 mm. If the outer diameter of the ring hole 4a is less than the above lower limit, it may be difficult for the direct light from the ring hole 4a to pass through the phase adjustment portion 9a of the phase shift plate 9, which will be described later. Conversely, if the outer diameter of the ring hole 4a exceeds the above upper limit, it may be difficult for the condenser lens 5 to collect the light.

[0039] The lower limit of the inner diameter of the ring hole 4a is preferably 1.3 mm, more preferably 1.5 mm. On the other hand, the upper limit of the inner diameter of the ring hole 4a is preferably 4 mm, more preferably 3.5 mm. If the inner diameter of the ring hole 4a is less than the above lower limit, it may be difficult for the direct light from the ring hole 4a to pass through the phase adjustment portion 9a of the phase shift plate 9, which will be described later. Conversely, if the inner diameter of the ring hole 4a exceeds the above upper limit, it may be difficult for the condenser lens 5 to focus the light.

[0040] The lower limit of the width of the ring hole 4a (the difference between the outer diameter and the inner diameter) is preferably 0.1 mm, more preferably 0.2 mm. On the other hand, the upper limit of the width of the ring hole 4a is preferably 0.5 mm, more preferably 0.4 mm. If the width of the ring hole 4a is less than the above lower limit, the amount of light may be insufficient, making it difficult to observe crystal defects. Conversely, if the width of the ring hole 4a exceeds the above upper limit, the image resolution may decrease, making it difficult to observe crystal defects.

[0041] <Condenser lens> The condenser lens 5 condenses the light emitted from the light source 1 that has passed through the ring diaphragm 4, and irradiates one surface of the semiconductor wafer W with the collected light.

[0042] The condenser lens 5 is appropriately selected depending on the numerical aperture of the objective lens selected in the objective lens unit 8, which will be described later. In other words, when a low-magnification objective lens 8a is selected in the objective lens unit 8, a different type of lens is used depending on whether a high-magnification objective lens 8b is selected.

[0043] The method for replacing the condenser lens 5 is not particularly limited, and the lens may be replaced each time, or a lens may be selected from a plurality of lenses mounted as a revolver, similar to the objective lens unit 8 described later.

[0044] The focal depth of the condenser lens 5 within the semiconductor wafer W is appropriately selected depending on the combination with the focal depth of the objective lens and the type of crystal defect. For example, when observing the epitaxial layer of the semiconductor wafer W, dislocations, which are crystal defects, have relatively small variations and fluctuations in the thickness direction of the semiconductor wafer W. Therefore, the lower limit of the focal depth of the condenser lens 5 is preferably 0.1 μm, more preferably 0.2 μm. On the other hand, the upper limit of the focal depth of the condenser lens 5 is preferably 2 μm, more preferably 1.5 μm. When observing the bulk crystal of the semiconductor wafer W, dislocations have relatively large variations and fluctuations in the thickness direction of the semiconductor wafer W. Therefore, the lower limit of the focal depth of the condenser lens 5 is preferably 2 μm, more preferably 4 μm. On the other hand, the upper limit of the focal depth of the condenser lens 5 is preferably 10 μm, more preferably 8 μm. In either case, if the focal depth is less than the lower limit, it may be difficult to capture crystal defects present deeper than the surface of the semiconductor wafer W. Conversely, if the depth of focus exceeds the upper limit, crystal defects may not be observed sufficiently.

[0045] If a moving mechanism (XY stage 6) described below is provided, this moving mechanism can easily capture crystal defects that exist at positions deeper than the surface of the semiconductor wafer W. Therefore, if the moving mechanism is provided, the focal depth of the condenser lens 5 within the semiconductor wafer W can be set to 0.1 μm or more and 0.5 μm or less.

[0046] <XYステージ> The XY stage 6 is a stand that supports the semiconductor wafer W, and is configured to be movable in the XY directions as well as in the direction of the central axis M.

[0047] The XY stage 6 can move in the XY direction to change the position of the semiconductor wafer W onto which the light emitted from the light source 1 is irradiated, so that the crystal defect observation device can comprehensively observe the entire surface of the semiconductor wafer W.

[0048] The XY stage 6 can also move in the direction of the central axis M. By linking it with an autofocus unit 7, which will be described later, the surface of the semiconductor wafer W can be easily positioned at a position where the light beam from the condenser lens 5 is focused.

[0049] Furthermore, the XY stage 6 can move in the direction of the central axis M independently of the condenser lens 5 and the objective lens selected by the objective lens unit 8. Therefore, for example, even if the XY stage 6 is moved in a direction approaching the objective lens, the distance between the objective lens and the condenser lens 5 is maintained constant. Viewed from the XY stage 6 side, when the objective lens moves relatively closer to the semiconductor wafer W in the optical axis direction (direction of the central axis M) (the distance between the XY stage 6 and the objective lens decreases), the condenser lens 5 moves relatively farther away from the semiconductor wafer W in the optical axis direction in synchronization with this, thereby maintaining the distance between the objective lens and the condenser lens 5 constant. In other words, the XY stage 6 functions as a movement mechanism that moves the condenser lens 5 relatively in the optical axis direction in synchronization with the relative movement of the objective lens in the optical axis direction (direction of the central axis M) with respect to the semiconductor wafer W while maintaining the distance between the objective lens and the condenser lens 5 constant. Here, "relative movement with respect to the semiconductor wafer" is a concept that includes not only moving the objective lens or condenser lens while the semiconductor wafer is fixed, but also moving the semiconductor wafer while the objective lens and condenser lens are fixed.

[0050] In this way, even if the XY stage 6 is moved in the optical axis direction while maintaining a constant distance between the objective lens and the condenser lens 5, the focal position does not change. When the autofocus unit 7 positions the focal position on the surface of the semiconductor wafer W, moving the XY stage 6 toward the objective lens moves the focal position toward the interior of the semiconductor wafer W, making it possible to observe the interior of the semiconductor wafer W. Because the optical path length passing through the interior of the semiconductor wafer W is relatively long, it is necessary to observe not only the surface but also the interior. In such cases, providing the XY stage 6 (moving mechanism) makes it easy to observe crystal defects located deep in the crystal axis direction.

[0051] <Autofocus unit> The autofocus unit 7 positions the semiconductor wafer W so that the light beams from the condenser lens 5 are focused on the surface of the semiconductor wafer W. Specifically, the autofocus unit 7 adjusts the XY stage 6 so that the position in the optical axis direction at which the light beams from the condenser lens 5 are focused coincides with the position of the surface of the semiconductor wafer W in the optical axis direction.

[0052] It is difficult to find the surface of the semiconductor wafer W, which is transparent with the naked eye using the visible light used for observation. Furthermore, since the focal depth of the crystal defect observation device is several μm to several tens of μm, it is not easy to accurately focus the light beam from the condenser lens 5 for such a focal depth. Therefore, by providing such an autofocus unit 7, the light beam from the condenser lens 5 can be easily focused on the surface of the semiconductor wafer W, thereby improving workability.

[0053] The autofocus unit 7 is not particularly limited, and a known autofocus mechanism using a laser can be used. Note that the wavelength of the laser light should be different from that of the light source 1 to prevent interference between them. For example, if the light source 1 has a wavelength of 500 nm, the wavelength of the laser light should be in the near-infrared region.

[0054] Furthermore, when the semiconductor wafer W is moved in the XY directions by the XY stage 6, warpage of the semiconductor wafer W may cause the surface of the semiconductor wafer W to deviate from the focusing position of the condenser lens 5. In such a case, by providing the autofocus unit 7, it is possible to detect the warpage of the semiconductor wafer W and adjust the position of the semiconductor wafer W in the optical axis direction according to the warpage, thereby preventing the surface of the semiconductor wafer W from deviating from the focusing position of the condenser lens 5. Therefore, even if the semiconductor wafer W is warped, it is possible to observe crystal defects without blurring the image.

[0055] <Objective lens unit> The objective lens unit 8 collects the transmitted light, which is transmitted through the semiconductor wafer W and emerges from the other surface, by the selected objective lens, and converts it into a parallel light beam again.

[0056] The objective lens selected by the objective lens unit 8 should be selected depending on the type of crystal defect. That is, the objective lens unit 8 should have a low-magnification objective lens 8a, a high-magnification objective lens 8b, and a revolver 8c for switching between them.

[0057] (low magnification objective lens) The lower limit of the numerical aperture of the low-magnification objective lens 8a is preferably 0.05, more preferably 0.1. On the other hand, the upper limit of the numerical aperture of the low-magnification objective lens 8a is preferably 0.15, more preferably 0.14. If the numerical aperture of the low-magnification objective lens 8a is below the lower limit, the brightness of the obtained image may decrease, and crystal defects may not be observed sufficiently. Conversely, if the numerical aperture of the low-magnification objective lens 8a is above the upper limit, the resolution may be insufficient, and crystal defects may not be observed sufficiently.

[0058] When using a low-magnification objective lens 8a having a numerical aperture within the above range, the lower limit of the numerical aperture of the condenser lens 5 is preferably 0.1, more preferably 0.2. On the other hand, the upper limit of the numerical aperture of the condenser lens 5 is preferably 0.5, more preferably 0.4. By setting the numerical apertures of the condenser lens and the objective lens within the above range, crystal defects in which dislocations are not aligned along the crystal axes and have significant fluctuations relative to the crystal axes can be suitably observed. Examples of crystal defects in which dislocations are not aligned along the crystal axes and have significant fluctuations relative to the crystal axes include crystal defects in bulk crystals.

[0059] The lower limit of the magnification of the low-magnification objective lens 8a is preferably 2x, more preferably 3x. On the other hand, the upper limit of the magnification of the low-magnification objective lens 8a is preferably 6x, more preferably 5x. If the magnification of the low-magnification objective lens 8a is below the lower limit, crystal defects may not be observed sufficiently. Conversely, if the magnification of the low-magnification objective lens 8a exceeds the upper limit, it may be difficult to capture crystal defects that exist deeper than the surface of the semiconductor wafer W.

[0060] The lower limit of the focal depth of the low-magnification objective lens 8a within the semiconductor wafer W is preferably 5 μm, more preferably 8 μm. On the other hand, the upper limit of the focal depth is preferably 15 μm, more preferably 12 μm. If the focal depth is less than the lower limit, it may be difficult to capture crystal defects that exist deeper than the surface of the semiconductor wafer W. Conversely, if the focal depth exceeds the upper limit, it may be difficult to observe the crystal defects sufficiently.

[0061] The upper limit of the resolution of the low-magnification objective lens 8a is preferably 3 μm, more preferably 2.5 μm. If the resolution exceeds the upper limit, crystal defects may not be observed sufficiently. On the other hand, the smaller the resolution of the low-magnification objective lens 8a, the better, but from the viewpoint of the cost of the low-magnification objective lens 8a, it is usually set to 1.5 μm or more.

[0062] (high magnification objective lens) The lower limit of the numerical aperture of the high-magnification objective lens 8b is preferably 0.2, more preferably 0.25. On the other hand, the upper limit of the numerical aperture of the high-magnification objective lens 8b is preferably 0.4, more preferably 0.35. If the numerical aperture of the high-magnification objective lens 8b is below the lower limit, the brightness of the obtained image may decrease, and crystal defects may not be observed sufficiently. Conversely, if the numerical aperture of the high-magnification objective lens 8b exceeds the upper limit, the resolution may be insufficient, and crystal defects may not be observed sufficiently.

[0063] When using a high-magnification objective lens 8b having a numerical aperture within the above range, the lower limit of the numerical aperture of the condenser lens 5 is preferably 0.6, more preferably 0.7. On the other hand, the upper limit of the numerical aperture of the condenser lens 5 is preferably 1, more preferably 0.95. By setting the numerical apertures of the condenser lens and the objective lens within the above ranges, crystal defects in which dislocations are aligned along the crystal axis can be suitably observed. Examples of crystal defects in which dislocations are aligned along the crystal axis include crystal defects in epitaxial layers.

[0064] The lower limit of the magnification of the high-magnification objective lens 8b is preferably 8x, and more preferably 9x. On the other hand, the upper limit of the magnification of the high-magnification objective lens 8b is preferably 20x, and more preferably 15x. If the magnification of the high-magnification objective lens 8b is below the lower limit, crystal defects may not be observed sufficiently. Conversely, if the magnification of the high-magnification objective lens 8b exceeds the upper limit, it may be difficult to capture crystal defects that exist deeper than the surface of the semiconductor wafer W.

[0065] The lower limit of the focal depth of the high-magnification objective lens 8b within the semiconductor wafer W is preferably 1 μm, more preferably 2 μm. On the other hand, the upper limit of the focal depth is preferably 10 μm, more preferably 5 μm. If the focal depth is less than the lower limit, it may be difficult to capture crystal defects that exist deeper than the surface of the semiconductor wafer W. Conversely, if the focal depth exceeds the upper limit, it may be difficult to observe the crystal defects sufficiently.

[0066] The upper limit of the resolution of the high-magnification objective lens 8b is preferably 2 μm, more preferably 1.5 μm. If the resolution exceeds the upper limit, crystal defects may not be observed sufficiently. On the other hand, the smaller the resolution of the high-magnification objective lens 8b, the better. However, from the viewpoint of the cost of the high-magnification objective lens 8b, the resolution is usually set to 0.5 μm or more.

[0067] (Nosepiece) The revolver 8c is a rotatable component to which multiple objective lenses (low-magnification objective lens 8a and high-magnification objective lens 8b) can be attached. By rotating the revolver 8c, one objective lens can be selected and positioned on the optical axis.

[0068] <Phase shift plate> The phase shift plate 9 is disposed at a position optically conjugate with the ring diaphragm 4. Specifically, it is disposed between an objective lens unit 8 having an objective lens and an imaging lens 10. The phase shift plate 9 may be ring-shaped.

[0069] The transparent or semi-transparent phase shift plate 9 generates a phase difference between a ring-shaped portion through which direct light passes and another portion through which diffracted light passes, of the transmitted light that has passed through the semiconductor wafer W. Specifically, as shown in Fig. 3, the phase shift plate 9 has a phase adjustment portion 9a, which is the ring-shaped portion through which direct light passes and which changes the phase of the direct light, and a passing portion 9b, which is the other portion through which diffracted light passes and which does not change the phase of the diffracted light.

[0070] It is preferable that the phase adjustment unit 9a can be aligned so as to overlap with the image of the aperture ring 4. In this case, the phase adjustment unit 9a is ring-shaped. If the phase adjustment unit 9a is not ring-shaped, there will be a portion that does not overlap with the image of the aperture ring 4, which may weaken the phase contrast.

[0071] The phase adjustment amount of the phase adjustment unit 9a is not particularly limited, but is preferably 1 / 4λ. Also, it is preferable to reduce the light intensity by the phase adjustment unit 9a. By adjusting the phase and light intensity by the phase adjustment unit 9a in this way, it becomes possible to capture crystal defects as an image.

[0072] The principle by which an image of a crystal defect can be obtained by passing the light through the phase shift plate 9 in the crystal defect observation device will be briefly explained with reference to FIG.

[0073] The light emitted from the light source 1 passes through the field diaphragm 3 where it is narrowed down, and then passes through the ring diaphragm 4 as shown in Fig. 4. In the ring diaphragm 4, the light can only pass through the ring hole 4a, so the light is incident in the form of rays onto the condenser lens 5 as shown in Fig. 4. The condenser lens 5 focuses the incident rays onto the surface (or inside) of the semiconductor wafer W.

[0074] If a crystal defect exists at the light-focus position, the crystal defect has a different refractive index from the surrounding area, and so the light splits into a direct light beam (solid line in Figure 4) that travels straight and a diffracted light beam (dashed line in Figure 4). This direct light beam and diffracted light beam are imaged again on the screen of the camera 11 by the objective lens (low-magnification objective lens 8a or high-magnification objective lens 8b). At this image-formed position, Imaging light = direct light + diffracted light The relationship is established.

[0075] As described above, the phase shift plate 9 is placed between the objective lens and the imaging position, with the direct light passing through the phase adjustment section 9a and the diffracted light passing through the passing section 9b. Conversely, the phase shift plate 9 is designed and positioned so that the direct light passes through the phase adjustment section 9a and the diffracted light passes through the passing section 9b.

[0076] Without this phase shift plate 9, neither the direct light nor the diffracted light undergoes any change. In this case, the brightness of the light that passes through the crystal defect and the light that does not pass through the crystal defect remains unchanged and cannot be distinguished, so the crystal defect cannot be observed from the formed image.

[0077] However, when the phase shift plate 9 is present, as described above, only the direct light is shifted in phase and dimmed, so the direct light and the diffracted light interfere with each other, causing the image-forming light to lose its original brightness and become dark. On the other hand, if there are no crystal defects, there is no diffracted light, so even if the phase, etc. changes, there is no interference and the brightness remains the same. In other words, because the brightness changes only in the areas where crystal defects exist, the crystal defect observation device can directly observe the crystal defects.

[0078] <Imaging lens> The imaging lens 10 forms an image of the transmitted light that has passed through the phase shift plate 9. Specifically, the imaging lens 10 is disposed downstream of the phase shift plate 9 in the direction of the transmitted light.

[0079] <Camera> The camera 11 is a tool for observing the image formed according to the above-mentioned principle. The camera 11 is not particularly limited, and any known camera can be used.

[0080] <Polarizer> The polarizer 12 is disposed between the light source 1 and the condenser lens 5. In the crystal defect observation device shown in FIG.

[0081] Polarizer 12 converts the light emitted from light source 1 from natural light into linearly polarized light. In other words, the light that passes through polarizer 12 becomes a transverse wave that vibrates only in one plane (polarization plane) perpendicular to the traveling direction of the light.

[0082] As the polarizer 12, a known film polarizer, prism polarizer, or the like can be used.

[0083] <Analyzer> The analyzer 13 is disposed between the phase shift plate 9 and the imaging lens 10. By disposing the polarizer 12 and the analyzer 13 in this manner, the polarization of the light used in the phase contrast observation is the same, which makes it easier to perform stable phase contrast observation.

[0084] The analyzer 13 transmits only the transverse wave component vibrating on one plane (polarization plane) perpendicular to the traveling direction of the light, out of the light that has passed through the phase shift plate 9. The analyzer 13 can have a configuration similar to that of the polarizer 12.

[0085] In this crystal defect observation device, the polarization relationship between the polarizer 12 and the analyzer 13 can be switched between parallel polarization (see FIG. 5) and orthogonal polarization (see FIG. 6). Here, in the case of parallel polarization, the polarization planes of the polarizer 12 and the analyzer 13 are on the same plane. On the other hand, in the case of orthogonal polarization, the polarization planes of the polarizer 12 and the analyzer 13 are orthogonal.

[0086] In this crystal defect observation device, the light irradiated onto the semiconductor wafer W is a transverse wave that vibrates only on one plane (polarization plane) perpendicular to the traveling direction of the light. Therefore, as shown in Fig. 5, when the polarizer 12 and the analyzer 13 are in a parallel polarization relationship, most of the light that has passed through the semiconductor wafer W passes through the analyzer 13. In this case, both the light that has passed through the screw dislocation and the light that has passed through the edge dislocation can pass through the analyzer 13, making it possible to observe both types of dislocations.

[0087] On the other hand, as shown in Figure 6, when the polarizer 12 and analyzer 13 are in an orthogonal polarization relationship, most of the light does not pass through the analyzer 13. The only light that can pass through the analyzer 13 is a component whose polarization direction is modulated by crystal defects present in the semiconductor wafer W. According to the observations of the present inventors, the polarization direction is modulated when light passes through a screw dislocation, but it appears that the polarization direction is not modulated when light passes through an edge dislocation. For this reason, when the polarizer 12 and analyzer 13 are orthogonally polarized, screw dislocations are mainly selectively observed.

[0088] Therefore, by making it possible to switch the polarization relationship between parallel polarization and orthogonal polarization, it is possible to easily classify the crystal defects of the semiconductor wafer W into screw dislocations and edge dislocations.

[0089] When switching the polarization relationship, the polarizer 12 alone or both the polarizer 12 and the analyzer 13 may be rotated. However, as shown in FIGS. 5 and 6, it is preferable to rotate the analyzer 13 alone. In this case, the polarizer 12 does not rotate, so that the semiconductor wafer W is always irradiated with transverse waves oscillating in the same polarization plane, enabling stable inspection. On the other hand, when the polarizer 12 is rotated to switch, transverse waves oscillating in different polarization planes are irradiated onto the semiconductor wafer W. Since the light irradiated onto the semiconductor wafer W contains components that are reflected within the semiconductor wafer W, when transverse waves oscillating in different polarization planes are irradiated, the transverse waves oscillating in different polarization planes coexist on the semiconductor wafer W and are output from the semiconductor wafer W, especially for a while after the polarization relationship is switched. In this case, the light output from the semiconductor wafer W contains transverse waves oscillating in unintended polarization planes, which may reduce the observation accuracy.

[0090] <Advantages> The crystal defect observation device directly observes changes in the refractive index of crystal defects by observing the phase difference of light passing through the semiconductor wafer W. Therefore, by using the crystal defect observation device, it is possible to accurately observe the crystal defects of even a transparent semiconductor wafer W. Furthermore, the crystal defect observation device can extract only specific crystal defects by setting the polarization relationship between the polarizer 12 and the analyzer 13 to a specific relationship, so that the types of crystal defects in the semiconductor wafer W can be easily classified.

[0091] [Other embodiments] The present disclosure is not limited to the above-described embodiments, and can be implemented in various forms including the above-described embodiments, as well as forms with various modifications and improvements.

[0092] In the above embodiment, the crystal defect observation device was described as moving the XY stage in the optical axis direction to observe the inside of a semiconductor wafer, but it is also possible to observe the inside of a semiconductor wafer by fixing the XY stage and moving the objective lens and condenser lens in the optical axis direction while maintaining the distance between them.

[0093] In the above embodiment, the crystal defect observation device is described as being equipped with an autofocus unit, but the autofocus unit is not a required component, and the present disclosure also intends to employ a crystal defect observation device that does not have an autofocus unit.

[0094] In the above embodiment, the phase shift plate of the crystal defect observation device is configured to adjust the phase in the portion through which direct light passes, but it is also possible to adjust the phase in the portion through which diffracted light passes. In this case, the phase adjustment is not performed in the portion through which direct light passes.

[0095] In the above embodiment, the phase shift plate is provided independently, but the phase shift plate may be formed as a ring-shaped phase shift film by vapor deposition or sputtering on a transparent glass plate inside the objective lens.

[0096] In the above embodiment, the crystal defect observation device is described as being equipped with a camera, but the method of observing crystal defects is not limited to a camera, and other recording devices such as a video recorder may also be used, or display devices such as a display may also be used.

[0097] In the above embodiment, the polarizer is disposed between the field stop and the ring stop, but the polarizer may be disposed between the light source and the field stop, or may be disposed at another position between the light source and the semiconductor wafer.

[0098] In the above embodiment, the analyzer is disposed between the phase shift plate and the imaging position, but it may be disposed at another position between the semiconductor wafer and the imaging position.

[0099] In the above embodiment, the polarization relationship between the polarizer and the analyzer is described as being switchable between parallel and orthogonal polarization. However, the polarization relationship between the polarizer and the analyzer is not limited to this relationship. For example, the device may be used solely for observing screw dislocations using orthogonal polarization. Furthermore, the device may be configured to further classify dislocations by combining it with, for example, circular polarization. Specifically, the present inventors have confirmed that dislocations that were not visible with parallel and orthogonal polarization become visible when combined with circular polarization, and have determined that the Burgers vectors of individual screw dislocations can be identified. [Industrial Applicability]

[0100] The crystal defect observation device of the present disclosure can easily classify the types of crystal defects in semiconductor wafers. [Explanation of symbols]

[0101] 1 light source 2 reflector 3 Field diaphragm 4 Aperture ring 4a Ring hole 4b Support girder 5. Condenser Lens 6 XY stages 7 Autofocus Unit 8 Objective Lens Unit 8a Low magnification objective lens 8b High-magnification objective lens 8c revolver 9 Phase shift plate 9a Phase adjustment section 9b Passage section 10 Imaging lens 11 Camera 12 Polarizer 13 Analyzer W semiconductor wafer M center axis

Claims

1. A crystal defect observation device for observing crystal defects in a semiconductor wafer, comprising: a light source collimated into a parallel beam; a ring diaphragm that confines light emitted from the light source into a ring shape; a condenser lens that condenses the emitted light that has passed through the ring diaphragm and irradiates the light onto one surface of the semiconductor wafer; an objective lens that condenses transmitted light that passes through the semiconductor wafer and emerges from the other surface to form a parallel beam again; a phase shift plate disposed at an optically conjugate position with the annular diaphragm; an imaging lens that forms an image of the transmitted light that has passed through the phase shift plate; a polarizer disposed between the light source and the condenser lens; an analyzer disposed between the phase shift plate and the imaging lens; Equipped with The phase shift plate generates a phase difference between a ring-shaped portion of the transmitted light through which direct light passes and another portion through which diffracted light passes.

2. 2. The crystal defect observation device according to claim 1, wherein the polarizer is disposed between the light source and the annular aperture.

3. 3. The crystal defect observation device according to claim 1, wherein the polarization relationship between said polarizer and said analyzer is switchable between parallel polarization and orthogonal polarization.

Images