JP7832558B2 - Waveguide device inspection system - Google Patents

Description

This disclosure relates to a waveguide device inspection system.

With the advancement of optical fiber communication, the practical application of optical communication devices such as optical transceivers and optical switches is progressing. These optical devices are formed on wafers made of silicon or compound semiconductors. This is because it allows for the effective use of technologies cultivated in the manufacturing of electronic devices, such as photolithography and dry etching, and offers excellent mass-production capabilities.

Optical devices handle optical signals with wavelengths (approximately 1 μm), which are significantly longer than the wavelengths of electrons (approximately 0.1 nm, depending on the velocity). Therefore, the signal transmission path (i.e., waveguide) cannot be bent sharply. Consequently, the paths in optical devices are significantly coarser than those in electronic devices, and the area occupied by the circuit pattern is also much larger in optical devices (on the order of ^cm² ) compared to electronic devices (on the order of ^mm² ). This has led to a problem where the computational cost required for automated visual inspection during manufacturing of optical devices is an order of magnitude higher than that of electronic devices, resulting in a deterioration of inspection throughput.

Furthermore, due to differences in the wavelengths of the elementary particles responsible for signal transport, the required performance differs even in the visual inspection of wafers manufactured for electronic and optical devices. In particular, in waveguide devices where waveguide patterns are formed, high precision is required in the inspection of the waveguide pattern, including the evanescent field, which is a minute region.

Optical signals propagating along a waveguide pattern are confined within the core region by total internal reflection at the core-cladding interface. However, the optical electric field of the optical signal seeps into the cladding layer, and this seeping electric field is called the evanescent field. The penetration length of the evanescent field into the cladding layer (hereinafter referred to as the evanescent length) is usually several micrometers to tens of micrometers, depending on the refractive index difference between the core and cladding. Foreign matter and defects present in the waveguide pattern region containing this evanescent field affect the optical properties of the waveguide device (such as the propagation characteristics of optical signals). Therefore, visual inspection of the waveguide pattern region containing this evanescent field requires high judgment accuracy.

In the visual inspection of waveguide patterns during the manufacturing of such waveguide device wafers, it is conceivable to determine acceptance or rejection by visual inspection with the naked eye. However, visual inspection with the naked eye results in low accuracy in inspecting even minute areas such as the evanescent field, and costs may increase as process operations increase.

Japanese Patent Application Publication No. 3-026945

M. Ota, K. Yamaguchi, and K. Suzuki, “Generative-adversarial-network-based dimensional measurement of optical waveguides,” Opt. Express 30, 6365-6373 (2022).

This disclosure has been made in view of the above-mentioned problems, and its purpose is to provide a waveguide device inspection system that enables highly accurate and efficient inspection of foreign matter and defects present in the region of the waveguide pattern including the evanescent field in a waveguide device.

To address the above-mentioned challenges, this disclosure provides a waveguide device inspection system for determining inspection results using difference images, comprising: an imaging device for acquiring an image of the area under inspection; a storage device for storing the acquired image; and a computing device configured to generate a difference image by comparing the acquired image with a reference image previously acquired by the imaging device. The computing device is further configured to refer to a design image of a waveguide device on which the waveguide pattern formed by the object under inspection is formed, and to generate a difference image in which only the portion affecting the performance of the waveguide device is extracted. In extracting only the portion affecting the performance of the waveguide device, the width of the portion affecting the performance of the waveguide device is set using the effective waveguide width (We), which is the sum of the waveguide width (W) of the waveguide pattern and twice the evanescent length (d) (W + 2d).

This diagram schematically illustrates the manufacturing process of a wafer for an example optical device. Figure 2 is a block diagram showing an example of the configuration of the waveguide device inspection system 200 in the first and second embodiments of this disclosure. This figure illustrates an example of the process performed by the waveguide device inspection system 200 of the first embodiment, which detects foreign objects or defects that unexpectedly occur in the pattern from the image acquired in the photolithography pattern inspection (corresponding to step 4 in Figure 1) described in Figure 1. This is a top view conceptually illustrating the masking of a waveguide pattern when an effective waveguide width We, including the evanescent field, is applied. This diagram shows a flowchart illustrating the processes performed by the waveguide device inspection system 200 in the first embodiment of this disclosure. This figure illustrates, as an example, the process by which the waveguide device inspection system 200 detects foreign objects or defects in a second embodiment of the present disclosure. This is a flowchart of the inspection process performed by the waveguide device inspection system 200 in the second embodiment of the present disclosure.

Various embodiments of this disclosure are described in detail below with reference to the drawings. Identical or similar reference numerals indicate identical or similar elements, and redundant descriptions may be omitted. Materials and numerical values are illustrative and not intended to limit the technical scope of this disclosure. The following description is illustrative and may involve omitting or modifying some configurations, or implementing them with additional configurations, without departing from the spirit of one embodiment of this disclosure.

(First embodiment)
A first embodiment of the waveguide device inspection system according to this disclosure will be described in detail below with reference to the drawings.

Figure 1 is a schematic diagram illustrating the manufacturing process of an exemplary optical device wafer. Figure 1 specifically illustrates the manufacturing process of a silica waveguide. As shown in Figure 1, the exemplary manufacturing process of a silica waveguide includes: forming a glass film on a substrate 101 that will form the lower cladding 102 and core 103 of the waveguide (Step 1); applying a photoresist 104 onto the glass layer that will form the core 103 (Step 2); transferring a circuit pattern by photolithography (Step 3); and performing a photolithographic pattern inspection to check for foreign matter or defects in the circuit pattern formed by photolithography (Step 4).

In step 1, an additive such as germanium is added to the glass film that will become the core 103 so that its refractive index is slightly higher than that of the glass that will become the lower cladding 102. Also, in Figure 1, the area indicated by the circle in step 3 shows a view from above of the wafer in the state in which the circuit pattern has been formed in step 3. In Figure 1, as an example, a wafer to be inspected is shown in which four circuit patterns 105-1 to 105-4 are formed.

In step 4, the area to be inspected (the area from which images (reference images and inspected images) described later are acquired) may be the entirety of the multiple circuit patterns transferred onto the wafer (in the example shown in Figure 1, all four areas 106-1 to 106-4), or it may be a divided area (in the example shown in Figure 1, for example, each of the four areas 106-1 to 106-4). Images of the area to be inspected can be captured using an imaging device capable of observing the area to be inspected, such as a microscope or camera. Then, the presence or absence of foreign matter or defects on the wafer is determined using these captured images. The detected foreign matter or defects may be caused, for example, by dust mixed in during resist coating in step 2, or by dust adhering to the wafer during photolithography in step 3.

Figure 2 is a block diagram showing an example of the configuration of a waveguide device inspection system 200 in a first embodiment of the present disclosure. As shown in Figure 2, the waveguide device inspection system 200 used in step 4 includes an imaging device 201 that acquires an image of the area to be inspected (reference image and inspection image), a storage device 205 that stores the acquired image, and a computing device 202 configured to compare the acquired inspection image and the reference image to generate a difference image (inspection image), and is an inspection system that detects foreign objects and defects in the inspection object using the generated inspection image. In Figure 2, an example configuration is shown that includes a display device 203 for displaying the inspection image, reference image, etc.

The imaging device 201 captures a reference image and an image under inspection, and may be capable of observing the area under inspection (for example, a camera, a magnifying glass such as an optical microscope, etc.). The computing device 202 is a device that controls the waveguide device inspection system 200, and controls the imaging device 201 and the storage device 205, and performs various processes using the reference image and the image under inspection (such as generating a design image for masks, masking, and generating a difference image, as described later).

The storage device 205 includes a design image storage unit 206, a mask design image storage unit 207, an image under inspection storage unit 208, a reference image storage unit 209, and a differential image storage unit 210. The imaging device 201, the computing device 202, the display device 203, and the storage device 205 are communicated with each other by a bus 212. The computing device 202 includes one or more processing units, which can be, for example, a central processing unit (CPU), a microprocessor unit (MPU), a graphics processing unit (GPU), a CISC (Complex Instruction Set Computer) type CPU, or a RISC (Reduced Instruction Set Computer) type CPU. The display device 203 can be, for example, a liquid crystal display (LCD) or a plasma display (PDP). The storage device 204 can be a semiconductor memory, a hard disk, or the like.

Figure 3 illustrates an example of the processing performed by the waveguide device inspection system 200 of the first embodiment, which detects foreign objects or defects that unexpectedly occur in a pattern from an image acquired in the photolithography pattern inspection (corresponding to step 4 in Figure 1) described in Figure 1. Figure 3 illustrates an example where the four divided regions 106-1 to 106-4 in Figure 1 are used one by one as the image to be inspected. As shown in Figure 3, in the processing performed by the waveguide device inspection system 200 of this embodiment, first, a mask design image 301 is generated in order to mask the parts other than the pattern that affect the optical device characteristics. This mask design image 301 is generated using a design image created from design data that reflects the circuit design of the wafer to be inspected. The pixel calculation for generating the mask design image 301 from the design image is performed by setting all pixels corresponding to areas that do not affect the performance of the optical device to "0" (i.e., "false") and setting the other parts to "1" (i.e., "true") for each pixel of the design image. In other words, the mask is set to extract only the areas that are likely to affect the performance of the optical device.

In this embodiment, the pattern inspection of a photolithography process when forming a waveguide pattern is described as an example. In such cases, it is preferable to set the region that affects the performance of the optical device (which becomes "1" in the mask design image) to a width slightly wider than the waveguide width, taking into account the evanescent field described above. The setting of the masking region for this waveguide pattern will be described in detail later.

Reference image 302 can be, for example, an image of an optical device in an ideal state free of foreign matter or defects, as captured by the imaging device 201 after visually inspecting the image taken in step 4 shown in Figure 1. A masked reference image 303 is generated by performing a pixel-by-pixel multiplication operation on this reference image 302 with the previously generated mask design image 301.

Next, the images to be inspected, 304 and 307, are acquired. These images can be captured using the imaging device 201. Similar to the reference image 302, a pixel-by-pixel multiplication operation is performed on these images 304 and 307 with the mask design image 301 to generate masked images 305 and 308.

The integration process between the mask design image 301 and the reference image 302, or the images under inspection 304, 307, may be performed after aligning the mask design image 301 with the reference image 302, and the mask design image 301 with the images under inspection 304, 307. For this alignment, for example, a pattern matching method is used for the circuit pattern of the design image having the same coordinate system as the mask design image 301 and the circuit pattern of the reference image 302 or the images under inspection 304, 307 acquired from the imaging device 201. Specifically, the positional misalignment is detected by aligning the coordinate system of the reference image 302 or the images under inspection 304, 307 with the coordinate system of the design image so that the circuit patterns of the design image match.

Finally, the masked reference image 303 and the masked images under examination 305 and 307 are compared to obtain the examination images 306 and 309.

The first inspection image 304 is an image taken in step 4 shown in Figure 1 of a wafer containing foreign matter A that does not affect the performance of the optical device. The first inspection image 306 is obtained by comparing the masked inspection image 305, which is generated by integrating the first inspection image 304 and the mask design image 301, with the aforementioned masked reference image 303. In the example shown in Figure 3, the foreign matter A that does not affect the performance of the optical device is not detected in the first inspection image 306 because it is located in the masked area. For reference, the original circuit pattern is shown as a dashed line in the first inspection image 306, but the circuit pattern is not displayed in the actual inspection image.

On the other hand, the second inspection image 307 is an image of a wafer containing foreign matter B, which does not affect the performance of the optical device, and foreign matter C, which does affect the performance of the optical device, as well as defects. Similarly, a pixel-by-pixel multiplication operation is performed on the second inspection image 307 with the design image to generate a masked second inspection image 308. As with the first inspection image 306, the second inspection image 309 is obtained by comparing the second inspection image 308 with the masked reference image 303. For reference, the original circuit pattern is also shown in the second inspection image 309 with a dashed line. In the second inspection image 309, foreign matter B is located in the masked area and is therefore not detected. On the other hand, foreign matter C and defects are located in the unmasked area and are therefore detected. In this way, the optical device chip including the second inspection image 307 is appropriately judged as unacceptable because only foreign matter C and defects that affect the performance of the optical device are detected. On the other hand, foreign object B is determined to be a foreign object that does not affect the performance of the optical device.

As a comparison process between the masked reference image 303 and the masked first image to be inspected 305 or second image to be inspected 308, for example, the difference images may be generated as inspection images 306 and 309 by simply performing a pixel-by-pixel difference process on each image. Alternatively, a threshold may be set for these difference images, and the inspection images may be generated by setting pixels with values below the threshold to "0" and pixels with values above the threshold to "1". It is also possible to perform the same processing after pre-binarizing the images to be inspected 304 and 307.

As described above, the waveguide device inspection system 200 shown in Figure 3 includes masking areas that do not affect the performance of the optical device using a design image created from the design data of the wafer to be inspected. In the case of pattern inspection of the photolithography process that forms the core pattern of the waveguide, it is preferable to set the masking area to a width slightly wider than the waveguide width. Here, if the design data is a waveguide pattern, by pre-setting the refractive indices of the waveguide core and cladding, the effective waveguide width We, which includes the evanescent field of the optical electric field seeping into the cladding as described above, can be determined. When setting the masking area of the waveguide pattern (the area where the pixel value is "1" (i.e., "true")), applying this effective waveguide width We, which includes the evanescent field, allows for efficient masking of the waveguide pattern.

In particular, We in a weak waveguide with low propagation loss can be calculated using (Equation 1).

Here, W is the waveguide width in the design data, λ is the wavelength of the optical electric field, n _ef is the effective refractive index of the waveguide, and n _clad is the refractive index of the cladding. Also, σ = 0 when the polarization of the propagation mode is TE (Transverse Electric) mode, and σ = 1 when it is TM (Transverse Magnetic) mode.

Figure 4 is a conceptual top view showing the masking of a waveguide pattern when an effective waveguide width We including the evanescent field is applied, where (a) is the image before masking and (b) is the image after masking. As described above, the electric field of the optical signal propagating within the core 103 seeps out to the cladding side as an evanescent field, and if its length (evanescent length) is d, then the effective waveguide width We including the evanescent field corresponds to W + 2d. As described above, foreign matter and defects present in the core 103 and the evanescent field region affect the optical propagation characteristics, so when the waveguide width in the design data is W, by setting the width of the region that is 1 (i.e., "true") in the design image to We, efficient waveguide pattern masking becomes possible.

Next, with reference to Figure 5, an example of the processing flow performed by the waveguide device inspection system 200 in this embodiment will be described below.

Figure 5 is a flowchart showing the processes performed by the waveguide device inspection system 200 in the first embodiment of this disclosure. When performing wafer inspection (corresponding to step 4 in Figure 1), a design image is generated from the design data (S401), and a mask design image C is obtained by referring to the generated design image and setting the pixel values of the regions that affect the performance of the optical device to "1" (i.e., "true") and the pixels of the other regions to "0" (i.e., "false") (S402). At this time, the value of each pixel expressed by the x-y coordinates of the mask design image C is defined as C(x,y). In addition, the regions in which the pixel value is set to "1" (i.e., "true") in the waveguide pattern masking are set using an effective waveguide width We that includes the evanescent field.

Furthermore, from the images of the inspection area of several wafers to be inspected, the image of a good product that has been confirmed to be free of foreign objects or defects is designated as reference image B of the inspection area (S403). At this time, the value of each pixel expressed by the x-y coordinates of inspection image B is defined as B(x,y).

Next, integration is performed on each pixel of reference image B and mask design image C. By performing integration on each pixel of reference image B and mask design image C, reference image B is masked to generate a masked reference image (S404). Here, the value of each pixel expressed by the x,y coordinates of the masked reference image can be defined as B(x,y)∧C(x,y), using "∧" as a symbol to indicate the multiplication operation for each pixel of the image. To perform the inspection, the image to be inspected is acquired and designated as the inspected image A (S405). Note that the order in which the mask design image C is obtained (S401, S402), the reference image B is obtained (S403, S404), and the inspected image A is obtained (S405) may be reversed or performed in parallel.

Next, by performing integration processing on each pixel of the image under inspection A and the mask design image C, the image under inspection A is masked, and a masked image under inspection is generated (S406). Here again, in the masking of the waveguide pattern, the region in which the pixel value is set to "1" (i.e., "true") is set using the effective waveguide width We which includes the evanescent field. At this time, the image under inspection A is captured, for example, in step 4 of Figure 1, using the imaging device 201 of the waveguide device inspection system 200 of this embodiment. Here, if the value of each pixel of the image under inspection A is A(x,y), then the value of each pixel of the masked image under inspection can similarly be defined as A(x,y)∧C(x,y).

Finally, the difference between the masked image under test and the masked reference image, that is, the difference between B(x,y) ∧ C(x,y) and A(x,y) ∧ C(x,y), is taken to generate the test image, which is the difference image (S407).

In the inspection method performed by this waveguide device inspection system 200, the inspection result is determined using the inspection image. That is, a pass/fail determination is made based on whether or not foreign matter or defects exist in the inspection image, which is a difference image (S408). This pass/fail determination may also be made by calculating the area of foreign matter and defects present in the inspection image and determining whether or not foreign matter and defects exist whose area exceeds a predetermined threshold. This makes it possible to target only the parts that affect the performance of the optical device for inspection and to make a pass/fail determination for inspections targeting only the foreign matter and defects present in those parts.

In the above process, both the image under inspection A and the reference image B were masked using the mask design image C. However, the same inspection can be performed by masking only the image under inspection A and extracting the inspection image from the difference between the masked image under inspection and the reference image B.

As described above, the waveguide device inspection system 200 in this embodiment allows for inspection only of parts that affect the performance of the optical device, and enables pass/fail judgments to be made only for foreign matter and defects present in those parts. Therefore, the computational cost required for automated visual inspection during the manufacturing of optical devices is reduced, and the degradation of inspection throughput is mitigated.

Furthermore, in the waveguide device inspection system 200 of this embodiment, the region for masking the waveguide pattern is set using the effective waveguide width We, which includes the evanescent field. Therefore, it becomes possible to efficiently and accurately inspect the region of the waveguide pattern that includes the evanescent field.

(Second embodiment)
A first embodiment of the waveguide device inspection system according to this disclosure will be described in detail below with reference to the drawings. In this embodiment as well, the description will be based on the example of applying the waveguide device inspection system of this embodiment to the inspection of the photolithography process in the wafer manufacturing process shown in Figure 1. Similar to the first embodiment, the waveguide device inspection system of this embodiment has the same configuration as the waveguide device inspection system 200 shown in Figure 2. Also, similar to the first embodiment, the mask design image 301 and the reference image 302 are generated in advance in the manner shown in Figure 3. In addition, the area of the waveguide pattern to be masked is set using an effective waveguide width We that includes the evanescent field, which is also the same as in the first embodiment.

Figure 6 illustrates, as an example, the process of detecting foreign objects and defects by the waveguide device inspection system 200 in a second embodiment of the present disclosure. In this embodiment, unlike the first embodiment, comparison images 501 and 504 are first generated by comparing the images under inspection 304 and 307 captured by the imaging device 201 with the reference image 302. In the comparison process, pixel-by-pixel difference calculation may be used, as in the first embodiment, or binarization may be performed in advance. The comparison images 501 and 504 thus generated include foreign object A or foreign object B that does not affect the performance of the optical device. Next, masked comparison images 502 and 505 are generated by multiplying the comparison images 501 and 504 with the mask design image 301, and finally inspection images 503 and 506 are obtained.

When performing integration processing between comparison images 501 and 504 and the mask design image 301, similar to the first embodiment, the masked comparison images 502 and 505 may be obtained by first aligning the two images and then masking the comparison image. For alignment, for example, a pattern matching method is used for the circuit pattern of the design image having the same coordinate system as the mask design image 301 and the circuit pattern of the reference image 302 or the images under inspection 304 and 307 acquired from the imaging device 201. Specifically, the positional misalignment is detected by aligning the coordinate system of the reference image 302 or the images under inspection 304 and 307 with the coordinate system of the design image so that the circuit patterns of the design image match.

Note that, as with Figure 3, the original circuit patterns are shown as dotted lines in comparison images 501 and 504, the masked comparison images 502 and 505, and the inspection images 503 and 506 for reference, but the circuit patterns are not displayed in the actual images.

The first inspection image 304 is an image taken in step 4 shown in Figure 1 of a wafer containing foreign matter A that does not affect the performance of the optical device. The first comparison image 501, generated by comparing this first inspection image 304 with the reference image 302, contains foreign matter A that does not affect the performance of the optical device. However, in the first inspection image 503, generated by multiplying this first comparison image 501 with the mask design image 301, the foreign matter A that does not affect the performance of the optical device is masked and removed. Therefore, similar to the first embodiment, foreign matter A is not detected in the first inspection image 503.

Furthermore, in the second comparison image 504, generated by comparing the second inspection image 307, which is an image of EHA containing foreign matter B that does not affect the performance of the optical device, and foreign matter C and defects that do affect the performance of the optical device, with the reference image, foreign matter B, C and defects are present. However, in the second inspection image 506, generated by multiplying the second comparison image 504 with the mask design image 301, foreign matter B, which does not affect the performance of the optical device, is masked and removed. Therefore, only foreign matter C and defects that affect the performance of the optical device are ultimately detected. As a result, the waveguide device inspection system 200 in this embodiment can appropriately determine that the optical device chip including the second inspection image 307 is unacceptable.

Next, with reference to Figure 7, an example of the processing flow performed by the waveguide device inspection system 200 in this embodiment will be described below.

Figure 7 is a flowchart of the inspection process performed by the waveguide device inspection system 200 in the second embodiment of this disclosure. In this embodiment as in the first embodiment, prior to inspection, a design image is generated from design data (S601), and a mask design image C is obtained by referring to the generated design image (S602). In addition, from the images of the inspection area taken for several wafers to be inspected, an image of a good product that has been confirmed to be free of foreign matter or defects is designated as reference image B (S603).

Next, the image under inspection A is acquired using the imaging device 201 (S604). Subsequently, the image under inspection A and the reference image B are compared to generate a difference image D(x,y) = A(x,y) - B(x,y) (S605). As described in the first embodiment, for example, D(x,y) is the value of each pixel expressed by the x-y coordinates of the inspection image D. By calculating D(x,y) ∧ C(x,y), which is the integration process of the generated difference image D and the mask design image C, a masked difference image, i.e., the inspection image, is generated (S606). Here again, as described in the first embodiment, "∧" is used as a symbol to indicate the pixel-by-pixel integration operation of the image.

Furthermore, the order in which the steps of obtaining the mask design image C (S601, S602), obtaining the reference image B (S603), obtaining the image under inspection A (S604), and calculating the difference between the image under inspection and the reference image to generate the inspection image (S605) are executed may be reversed or performed in parallel.

In this embodiment, the inspection method performed by the waveguide device inspection system 200 determines the inspection result using a masked difference image. That is, a pass/fail determination is made based on whether or not foreign matter or defects exist in the inspection images 503 and 506, which are masked difference images (S607). This pass/fail determination may also be made by calculating the area of foreign matter and defects present in the inspection images 503 and 506 and determining whether or not foreign matter and defects exist whose area exceeds a predetermined threshold. This makes it possible to inspect only the parts that affect the performance of the optical device and to make a pass/fail determination based only on foreign matter and defects present in those parts.

As described above, this embodiment also allows for inspection of only the parts that affect the performance of the optical device, and enables pass/fail judgments to be made only for foreign matter and defects present in those parts. Therefore, the computational cost required for automated visual inspection during the manufacturing of optical devices is reduced, and the degradation of inspection throughput is mitigated.

Furthermore, in this embodiment as well, similar to the first embodiment, the waveguide device inspection system 200 sets the region of the waveguide pattern masking using an effective waveguide width We that includes the evanescent field. Therefore, it becomes possible to efficiently and accurately inspect the region of the waveguide pattern that includes the evanescent field.

In the first and second embodiments, the visual inspection of a waveguide core pattern generated by the photolithography process of the waveguide circuit pattern in Figure 1 was used as an example. However, the waveguide device inspection system according to this disclosure can also be used to inspect wafers with control wiring formed on the waveguide, such as after the processing of the waveguide layer or for optical switches and optical variable attenuators. In particular, when other patterns are formed in the layer below the pattern to be inspected, the irregularities of the layer below at that location may be reflected in the pattern to be inspected, causing the appearance of the upper layer pattern to be inspected to differ. In such cases, false detections may be reduced by masking the parts of the upper layer pattern that appear different.

As described above, the waveguide device inspection system according to this disclosure is characterized by its ability to efficiently detect foreign matter and defects by masking areas that do not affect the performance of optical devices. Furthermore, by setting the area of the waveguide pattern to be masked based on the effective waveguide width We, which takes the evanescent field into consideration, it becomes possible to efficiently and accurately inspect the area of the waveguide pattern that includes the evanescent field. Having these features, the waveguide device inspection system according to this disclosure is expected to be applied as a highly efficient and accurate inspection system during the manufacturing of waveguide devices.

Claims (4)

A waveguide device inspection system that determines inspection results using difference images,
An imaging device that acquires an image of the area to be inspected,
A storage device for storing the acquired image,
The system includes a computing device configured to generate a difference image by comparing the acquired image with a reference image previously acquired by the imaging device,
The computing device is further configured to refer to a design image of a waveguide device on which the waveguide pattern of the object to be inspected is formed, and to generate a difference image in which only the portion that affects the performance of the waveguide device is extracted.
In extracting only the portion that affects the performance of the waveguide device, the width of the portion that affects the performance of the waveguide device is set using the effective waveguide width (We), which is obtained by the sum of the waveguide width (W) of the waveguide pattern and twice the evanescent length (d) (W + 2d).
A waveguide device inspection system characterized by the following features. The computing device is further configured to generate a difference image in which only the portion affecting the performance of the waveguide device is extracted, using a mask design image extracted by referring to the design image.
The aforementioned mask design image has true or 1 values for the pixel coordinates in the portion that affects the waveguide device, and false or 0 values for the pixel coordinates in the other portions.
When the value of each pixel in the acquired image is denoted as A(x,y), the value of each pixel in the reference image is B(x,y), the value of each pixel in the mask design image is C(x,y), and the pixel-wise product operation of the images is denoted as ∧,
The aforementioned computing device is
The difference between the image generated by the operation A(x,y) ∧ C(x,y) and the reference image,
or
The difference between the image generated by the operation A(x,y) ∧ C(x,y) and the image generated by the operation B(x,y) ∧ C(x,y),
The waveguide device inspection system according to claim 1, characterized in that the difference image is generated by any of the following methods. The aforementioned computing device is
The system is further configured to generate a difference image in which only the portion affecting the performance of the waveguide device is extracted, using a mask design image extracted by referring to the aforementioned design image.
The aforementioned mask design image has a pixel coordinate value of true or 1 in the portion that affects the performance of the waveguide device, and a pixel coordinate value of false or 0 in the other portion.
When the values of each pixel in the difference image between the acquired image and the reference image are denoted as D(x, y), the values of each pixel in the mask design image are denoted as C(x, y), and the pixel-wise product operation of the images is denoted as ∧,
The waveguide device inspection system according to claim 1, characterized in that the calculation device generates the difference image by the calculation D(x,y) ∧ C(x,y). The waveguide device inspection system according to any one of claims 1 to 3, characterized in that the image of the area to be inspected is an image of a waveguide device formed on a wafer.